open-vault/builtin/logical/pki/backend_test.go

6094 lines
204 KiB
Go
Raw Normal View History

package pki
import (
"bytes"
"context"
"crypto"
"crypto/ecdsa"
"crypto/ed25519"
"crypto/elliptic"
"crypto/rand"
"crypto/rsa"
"crypto/x509"
"crypto/x509/pkix"
"encoding/base64"
"encoding/hex"
"encoding/json"
"encoding/pem"
"fmt"
"math"
"math/big"
mathrand "math/rand"
2015-10-14 15:46:01 +00:00
"net"
2018-06-15 19:32:25 +00:00
"net/url"
"os"
2015-10-14 15:46:01 +00:00
"reflect"
"sort"
"strconv"
"strings"
2018-02-20 05:03:45 +00:00
"sync"
"sync/atomic"
"testing"
"time"
"github.com/stretchr/testify/require"
"github.com/armon/go-metrics"
"github.com/fatih/structs"
"github.com/go-test/deep"
"github.com/hashicorp/go-secure-stdlib/strutil"
"github.com/hashicorp/vault/api"
auth "github.com/hashicorp/vault/api/auth/userpass"
"github.com/hashicorp/vault/builtin/credential/userpass"
logicaltest "github.com/hashicorp/vault/helper/testhelpers/logical"
vaulthttp "github.com/hashicorp/vault/http"
"github.com/hashicorp/vault/sdk/helper/certutil"
"github.com/hashicorp/vault/sdk/logical"
"github.com/hashicorp/vault/vault"
"github.com/mitchellh/mapstructure"
2018-02-10 15:07:10 +00:00
"golang.org/x/net/idna"
)
var stepCount = 0
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
// From builtin/credential/cert/test-fixtures/root/rootcacert.pem
const (
rootCACertPEM = `-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----`
rootCAKeyPEM = `-----BEGIN RSA PRIVATE KEY-----
MIIEpQIBAAKCAQEA7FMy+FO4hFzZiHFmmY/B6A/zpyCep9PWZfGLUDtDuprHbg2p
t/TQjegMxC0AmWPZEHeG9FIJvT+WQXuLUG5v5MbG4zs21mqnYXwWAbTPaZ35bpp3
BbbquXFtAqOZG8yfrob1g9OSgmY+bG3ZNxxv35qmbBbuWyUvmPwfjzEAclMxv48w
/2Rs4dXkBuvc9PiNdQ9Sv+ZYG8GIqIcbRd38cSaXI4Q94BOHEr4jm1vqb54H7twv
0Gy9CnLXfn7ZtEDFSmG8WLk2GvIna+UC+gnxSqKCw0rcTby63rQUCgdJZF8VQUVi
18+BMNLXI4pT/P9cxTaCJC/YeuV5a3SaniOoFQIDAQABAoIBAQCoGZJC84JnnIgb
ttZNWuWKBXbCJcDVDikOQJ9hBZbqsFg1X0CfGmQS3MHf9Ubc1Ro8zVjQh15oIEfn
8lIpdzTeXcpxLdiW8ix3ekVJF20F6pnXY8ZP6UnTeOwamXY6QPZAtb0D9UXcvY+f
nw+IVRD6082XS0Rmzu+peYWVXDy+FDN+HJRANBcdJZz8gOmNBIe0qDWx1b85d/s8
2Kk1Wwdss1IwAGeSddTSwzBNaaHdItZaMZOqPW1gRyBfVSkcUQIE6zn2RKw2b70t
grkIvyRcTdfmiKbqkkJ+eR+ITOUt0cBZSH4cDjlQA+r7hulvoBpQBRj068Toxkcc
bTagHaPBAoGBAPWPGVkHqhTbJ/DjmqDIStxby2M1fhhHt4xUGHinhUYjQjGOtDQ9
0mfaB7HObudRiSLydRAVGAHGyNJdQcTeFxeQbovwGiYKfZSA1IGpea7dTxPpGEdN
ksA0pzSp9MfKzX/MdLuAkEtO58aAg5YzsgX9hDNxo4MhH/gremZhEGZlAoGBAPZf
lqdYvAL0fjHGJ1FUEalhzGCGE9PH2iOqsxqLCXK7bDbzYSjvuiHkhYJHAOgVdiW1
lB34UHHYAqZ1VVoFqJ05gax6DE2+r7K5VV3FUCaC0Zm3pavxchU9R/TKP82xRrBj
AFWwdgDTxUyvQEmgPR9sqorftO71Iz2tiwyTpIfxAoGBAIhEMLzHFAse0rtKkrRG
ccR27BbRyHeQ1Lp6sFnEHKEfT8xQdI/I/snCpCJ3e/PBu2g5Q9z416mktiyGs8ib
thTNgYsGYnxZtfaCx2pssanoBcn2wBJRae5fSapf5gY49HDG9MBYR7qCvvvYtSzU
4yWP2ZzyotpRt3vwJKxLkN5BAoGAORHpZvhiDNkvxj3da7Rqpu7VleJZA2y+9hYb
iOF+HcqWhaAY+I+XcTRrTMM/zYLzLEcEeXDEyao86uwxCjpXVZw1kotvAC9UqbTO
tnr3VwRkoxPsV4kFYTAh0+1pnC8dbcxxDmhi3Uww3tOVs7hfkEDuvF6XnebA9A+Y
LyCgMzECgYEA6cCU8QODOivIKWFRXucvWckgE6MYDBaAwe6qcLsd1Q/gpE2e3yQc
4RB3bcyiPROLzMLlXFxf1vSNJQdIaVfrRv+zJeGIiivLPU8+Eq4Lrb+tl1LepcOX
OzQeADTSCn5VidOfjDkIst9UXjMlrFfV9/oJEw5Eiqa6lkNPCGDhfA8=
-----END RSA PRIVATE KEY-----`
)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
func TestPKI_RequireCN(t *testing.T) {
t.Parallel()
b, s := CreateBackendWithStorage(t)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err := CBWrite(b, s, "root/generate/internal", map[string]interface{}{
"common_name": "myvault.com",
})
if err != nil {
t.Fatal(err)
}
if resp == nil {
t.Fatal("expected ca info")
}
// Create a role which does require CN (default)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "roles/example", map[string]interface{}{
"allowed_domains": "foobar.com,zipzap.com,abc.com,xyz.com",
"allow_bare_domains": true,
"allow_subdomains": true,
"max_ttl": "2h",
})
if err != nil {
t.Fatal(err)
}
// Issue a cert with require_cn set to true and with common name supplied.
// It should succeed.
_, err = CBWrite(b, s, "issue/example", map[string]interface{}{
"common_name": "foobar.com",
})
if err != nil {
t.Fatal(err)
}
// Issue a cert with require_cn set to true and with out supplying the
// common name. It should error out.
_, err = CBWrite(b, s, "issue/example", map[string]interface{}{})
if err == nil {
t.Fatalf("expected an error due to missing common_name")
}
// Modify the role to make the common name optional
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "roles/example", map[string]interface{}{
"allowed_domains": "foobar.com,zipzap.com,abc.com,xyz.com",
"allow_bare_domains": true,
"allow_subdomains": true,
"max_ttl": "2h",
"require_cn": false,
})
if err != nil {
t.Fatal(err)
}
// Issue a cert with require_cn set to false and without supplying the
// common name. It should succeed.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "issue/example", map[string]interface{}{})
if err != nil {
t.Fatal(err)
}
if resp.Data["certificate"] == "" {
t.Fatalf("expected a cert to be generated")
}
// Issue a cert with require_cn set to false and with a common name. It
// should succeed.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "issue/example", map[string]interface{}{})
if err != nil {
t.Fatal(err)
}
if resp.Data["certificate"] == "" {
t.Fatalf("expected a cert to be generated")
}
}
func TestPKI_DeviceCert(t *testing.T) {
t.Parallel()
b, s := CreateBackendWithStorage(t)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err := CBWrite(b, s, "root/generate/internal", map[string]interface{}{
"common_name": "myvault.com",
"not_after": "9999-12-31T23:59:59Z",
"not_before_duration": "2h",
})
if err != nil {
t.Fatal(err)
}
if resp == nil {
t.Fatal("expected ca info")
}
var certBundle certutil.CertBundle
err = mapstructure.Decode(resp.Data, &certBundle)
if err != nil {
t.Fatal(err)
}
parsedCertBundle, err := certBundle.ToParsedCertBundle()
if err != nil {
t.Fatal(err)
}
cert := parsedCertBundle.Certificate
notAfter := cert.NotAfter.Format(time.RFC3339)
if notAfter != "9999-12-31T23:59:59Z" {
t.Fatalf("not after from certificate: %v is not matching with input parameter: %v", cert.NotAfter, "9999-12-31T23:59:59Z")
}
if math.Abs(float64(time.Now().Add(-2*time.Hour).Unix()-cert.NotBefore.Unix())) > 10 {
t.Fatalf("root/generate/internal did not properly set validity period (notBefore): was %v vs expected %v", cert.NotBefore, time.Now().Add(-2*time.Hour))
}
// Create a role which does require CN (default)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "roles/example", map[string]interface{}{
"allowed_domains": "foobar.com,zipzap.com,abc.com,xyz.com",
"allow_bare_domains": true,
"allow_subdomains": true,
"not_after": "9999-12-31T23:59:59Z",
})
if err != nil {
t.Fatal(err)
}
// Issue a cert with require_cn set to true and with common name supplied.
// It should succeed.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "issue/example", map[string]interface{}{
"common_name": "foobar.com",
})
if err != nil {
t.Fatal(err)
}
err = mapstructure.Decode(resp.Data, &certBundle)
if err != nil {
t.Fatal(err)
}
parsedCertBundle, err = certBundle.ToParsedCertBundle()
if err != nil {
t.Fatal(err)
}
cert = parsedCertBundle.Certificate
notAfter = cert.NotAfter.Format(time.RFC3339)
if notAfter != "9999-12-31T23:59:59Z" {
t.Fatal(fmt.Errorf("not after from certificate is not matching with input parameter"))
}
}
func TestBackend_InvalidParameter(t *testing.T) {
t.Parallel()
b, s := CreateBackendWithStorage(t)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err := CBWrite(b, s, "root/generate/internal", map[string]interface{}{
"common_name": "myvault.com",
"not_after": "9999-12-31T23:59:59Z",
"ttl": "25h",
})
if err == nil {
t.Fatal(err)
}
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "root/generate/internal", map[string]interface{}{
"common_name": "myvault.com",
"not_after": "9999-12-31T23:59:59",
})
if err == nil {
t.Fatal(err)
}
}
2022-01-27 18:06:34 +00:00
2015-10-14 15:46:01 +00:00
func TestBackend_CSRValues(t *testing.T) {
t.Parallel()
2018-02-20 05:03:45 +00:00
initTest.Do(setCerts)
b, _ := CreateBackendWithStorage(t)
2015-10-14 15:46:01 +00:00
testCase := logicaltest.TestCase{
2018-11-07 01:21:24 +00:00
LogicalBackend: b,
Steps: []logicaltest.TestStep{},
2015-10-14 15:46:01 +00:00
}
intdata := map[string]interface{}{}
reqdata := map[string]interface{}{}
testCase.Steps = append(testCase.Steps, generateCSRSteps(t, ecCACert, ecCAKey, intdata, reqdata)...)
logicaltest.Test(t, testCase)
}
func TestBackend_URLsCRUD(t *testing.T) {
t.Parallel()
2018-02-20 05:03:45 +00:00
initTest.Do(setCerts)
b, _ := CreateBackendWithStorage(t)
testCase := logicaltest.TestCase{
2018-11-07 01:21:24 +00:00
LogicalBackend: b,
Steps: []logicaltest.TestStep{},
}
intdata := map[string]interface{}{}
reqdata := map[string]interface{}{}
testCase.Steps = append(testCase.Steps, generateURLSteps(t, ecCACert, ecCAKey, intdata, reqdata)...)
logicaltest.Test(t, testCase)
}
// Generates and tests steps that walk through the various possibilities
// of role flags to ensure that they are properly restricted
func TestBackend_Roles(t *testing.T) {
t.Parallel()
cases := []struct {
name string
key, cert *string
useCSR bool
}{
{"RSA", &rsaCAKey, &rsaCACert, false},
{"RSACSR", &rsaCAKey, &rsaCACert, true},
{"EC", &ecCAKey, &ecCACert, false},
{"ECCSR", &ecCAKey, &ecCACert, true},
{"ED", &edCAKey, &edCACert, false},
{"EDCSR", &edCAKey, &edCACert, true},
}
for _, tc := range cases {
tc := tc
t.Run(tc.name, func(t *testing.T) {
initTest.Do(setCerts)
b, _ := CreateBackendWithStorage(t)
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
testCase := logicaltest.TestCase{
LogicalBackend: b,
Steps: []logicaltest.TestStep{
{
Operation: logical.UpdateOperation,
Path: "config/ca",
Data: map[string]interface{}{
"pem_bundle": *tc.key + "\n" + *tc.cert,
},
},
},
}
testCase.Steps = append(testCase.Steps, generateRoleSteps(t, tc.useCSR)...)
if len(os.Getenv("VAULT_VERBOSE_PKITESTS")) > 0 {
for i, v := range testCase.Steps {
data := map[string]interface{}{}
var keys []string
for k := range v.Data {
keys = append(keys, k)
}
sort.Strings(keys)
for _, k := range keys {
interf := v.Data[k]
switch v := interf.(type) {
case bool:
if !v {
continue
}
case int:
if v == 0 {
continue
}
case []string:
if len(v) == 0 {
continue
}
case string:
if v == "" {
continue
}
lines := strings.Split(v, "\n")
if len(lines) > 1 {
data[k] = lines[0] + " ... (truncated)"
continue
}
}
data[k] = interf
}
t.Logf("Step %d:\n%s %s err=%v %+v\n\n", i+1, v.Operation, v.Path, v.ErrorOk, data)
}
}
logicaltest.Test(t, testCase)
})
}
}
// Performs some validity checking on the returned bundles
func checkCertsAndPrivateKey(keyType string, key crypto.Signer, usage x509.KeyUsage, extUsage x509.ExtKeyUsage, validity time.Duration, certBundle *certutil.CertBundle) (*certutil.ParsedCertBundle, error) {
parsedCertBundle, err := certBundle.ToParsedCertBundle()
if err != nil {
return nil, fmt.Errorf("error parsing cert bundle: %s", err)
}
if key != nil {
switch keyType {
case "rsa":
parsedCertBundle.PrivateKeyType = certutil.RSAPrivateKey
parsedCertBundle.PrivateKey = key
parsedCertBundle.PrivateKeyBytes = x509.MarshalPKCS1PrivateKey(key.(*rsa.PrivateKey))
case "ec":
parsedCertBundle.PrivateKeyType = certutil.ECPrivateKey
parsedCertBundle.PrivateKey = key
parsedCertBundle.PrivateKeyBytes, err = x509.MarshalECPrivateKey(key.(*ecdsa.PrivateKey))
if err != nil {
return nil, fmt.Errorf("error parsing EC key: %s", err)
}
case "ed25519":
parsedCertBundle.PrivateKeyType = certutil.Ed25519PrivateKey
parsedCertBundle.PrivateKey = key
parsedCertBundle.PrivateKeyBytes, err = x509.MarshalPKCS8PrivateKey(key.(ed25519.PrivateKey))
if err != nil {
return nil, fmt.Errorf("error parsing Ed25519 key: %s", err)
}
}
}
switch {
case parsedCertBundle.Certificate == nil:
return nil, fmt.Errorf("did not find a certificate in the cert bundle")
case len(parsedCertBundle.CAChain) == 0 || parsedCertBundle.CAChain[0].Certificate == nil:
return nil, fmt.Errorf("did not find a CA in the cert bundle")
case parsedCertBundle.PrivateKey == nil:
return nil, fmt.Errorf("did not find a private key in the cert bundle")
case parsedCertBundle.PrivateKeyType == certutil.UnknownPrivateKey:
return nil, fmt.Errorf("could not figure out type of private key")
}
switch {
case parsedCertBundle.PrivateKeyType == certutil.Ed25519PrivateKey && keyType != "ed25519":
fallthrough
case parsedCertBundle.PrivateKeyType == certutil.RSAPrivateKey && keyType != "rsa":
fallthrough
case parsedCertBundle.PrivateKeyType == certutil.ECPrivateKey && keyType != "ec":
return nil, fmt.Errorf("given key type does not match type found in bundle")
}
cert := parsedCertBundle.Certificate
if usage != cert.KeyUsage {
return nil, fmt.Errorf("expected usage of %#v, got %#v; ext usage is %#v", usage, cert.KeyUsage, cert.ExtKeyUsage)
}
// There should only be one ext usage type, because only one is requested
// in the tests
if len(cert.ExtKeyUsage) != 1 {
return nil, fmt.Errorf("got wrong size key usage in generated cert; expected 1, values are %#v", cert.ExtKeyUsage)
}
switch extUsage {
case x509.ExtKeyUsageEmailProtection:
if cert.ExtKeyUsage[0] != x509.ExtKeyUsageEmailProtection {
return nil, fmt.Errorf("bad extended key usage")
}
case x509.ExtKeyUsageServerAuth:
if cert.ExtKeyUsage[0] != x509.ExtKeyUsageServerAuth {
return nil, fmt.Errorf("bad extended key usage")
}
case x509.ExtKeyUsageClientAuth:
if cert.ExtKeyUsage[0] != x509.ExtKeyUsageClientAuth {
return nil, fmt.Errorf("bad extended key usage")
}
case x509.ExtKeyUsageCodeSigning:
if cert.ExtKeyUsage[0] != x509.ExtKeyUsageCodeSigning {
return nil, fmt.Errorf("bad extended key usage")
}
}
// TODO: We incremented 20->25 due to CircleCI execution
// being slow and pausing this test. We might consider recording the
// actual issuance time of the cert and calculating the expected
// validity period +/- fuzz, but that'd require recording and passing
// through more information.
if math.Abs(float64(time.Now().Add(validity).Unix()-cert.NotAfter.Unix())) > 25 {
return nil, fmt.Errorf("certificate validity end: %s; expected within 25 seconds of %s", cert.NotAfter.Format(time.RFC3339), time.Now().Add(validity).Format(time.RFC3339))
}
return parsedCertBundle, nil
}
func generateURLSteps(t *testing.T, caCert, caKey string, intdata, reqdata map[string]interface{}) []logicaltest.TestStep {
expected := certutil.URLEntries{
IssuingCertificates: []string{
"http://example.com/ca1",
"http://example.com/ca2",
},
CRLDistributionPoints: []string{
"http://example.com/crl1",
"http://example.com/crl2",
},
OCSPServers: []string{
"http://example.com/ocsp1",
"http://example.com/ocsp2",
},
}
csrTemplate := x509.CertificateRequest{
Subject: pkix.Name{
CommonName: "my@example.com",
},
}
priv1024, _ := rsa.GenerateKey(rand.Reader, 1024)
csr1024, _ := x509.CreateCertificateRequest(rand.Reader, &csrTemplate, priv1024)
csrPem1024 := strings.TrimSpace(string(pem.EncodeToMemory(&pem.Block{
Type: "CERTIFICATE REQUEST",
Bytes: csr1024,
})))
priv2048, _ := rsa.GenerateKey(rand.Reader, 2048)
csr2048, _ := x509.CreateCertificateRequest(rand.Reader, &csrTemplate, priv2048)
csrPem2048 := strings.TrimSpace(string(pem.EncodeToMemory(&pem.Block{
Type: "CERTIFICATE REQUEST",
Bytes: csr2048,
})))
ret := []logicaltest.TestStep{
{
2016-01-07 15:30:47 +00:00
Operation: logical.UpdateOperation,
Path: "root/generate/exported",
Data: map[string]interface{}{
"common_name": "Root Cert",
"ttl": "180h",
},
Check: func(resp *logical.Response) error {
if resp.Secret != nil && resp.Secret.LeaseID != "" {
return fmt.Errorf("root returned with a lease")
}
return nil
},
},
{
2016-01-07 15:30:47 +00:00
Operation: logical.UpdateOperation,
Path: "config/urls",
Data: map[string]interface{}{
"issuing_certificates": strings.Join(expected.IssuingCertificates, ","),
"crl_distribution_points": strings.Join(expected.CRLDistributionPoints, ","),
"ocsp_servers": strings.Join(expected.OCSPServers, ","),
},
},
{
Operation: logical.ReadOperation,
Path: "config/urls",
Check: func(resp *logical.Response) error {
if resp.Data == nil {
return fmt.Errorf("no data returned")
}
var entries certutil.URLEntries
err := mapstructure.Decode(resp.Data, &entries)
if err != nil {
return err
}
if !reflect.DeepEqual(entries, expected) {
return fmt.Errorf("expected urls\n%#v\ndoes not match provided\n%#v\n", expected, entries)
}
return nil
},
},
{
2016-01-07 15:30:47 +00:00
Operation: logical.UpdateOperation,
Path: "root/sign-intermediate",
Data: map[string]interface{}{
2018-02-10 15:07:10 +00:00
"common_name": "intermediate.cert.com",
"csr": csrPem1024,
"format": "der",
},
ErrorOk: true,
Check: func(resp *logical.Response) error {
if !resp.IsError() {
return fmt.Errorf("expected an error response but did not get one")
}
if !strings.Contains(resp.Data["error"].(string), "2048") {
2018-03-20 18:54:10 +00:00
return fmt.Errorf("received an error but not about a 1024-bit key, error was: %s", resp.Data["error"].(string))
}
return nil
},
},
{
Operation: logical.UpdateOperation,
Path: "root/sign-intermediate",
Data: map[string]interface{}{
"common_name": "intermediate.cert.com",
"csr": csrPem2048,
"signature_bits": 512,
"format": "der",
"not_before_duration": "2h",
// Let's Encrypt -- R3 SKID
"skid": "14:2E:B3:17:B7:58:56:CB:AE:50:09:40:E6:1F:AF:9D:8B:14:C2:C6",
},
Check: func(resp *logical.Response) error {
certString := resp.Data["certificate"].(string)
if certString == "" {
return fmt.Errorf("no certificate returned")
}
if resp.Secret != nil && resp.Secret.LeaseID != "" {
return fmt.Errorf("signed intermediate returned with a lease")
}
certBytes, _ := base64.StdEncoding.DecodeString(certString)
certs, err := x509.ParseCertificates(certBytes)
if err != nil {
return fmt.Errorf("returned cert cannot be parsed: %w", err)
}
if len(certs) != 1 {
return fmt.Errorf("unexpected returned length of certificates: %d", len(certs))
}
cert := certs[0]
skid, _ := hex.DecodeString("142EB317B75856CBAE500940E61FAF9D8B14C2C6")
switch {
case !reflect.DeepEqual(expected.IssuingCertificates, cert.IssuingCertificateURL):
return fmt.Errorf("IssuingCertificateURL:\nexpected\n%#v\ngot\n%#v\n", expected.IssuingCertificates, cert.IssuingCertificateURL)
case !reflect.DeepEqual(expected.CRLDistributionPoints, cert.CRLDistributionPoints):
return fmt.Errorf("CRLDistributionPoints:\nexpected\n%#v\ngot\n%#v\n", expected.CRLDistributionPoints, cert.CRLDistributionPoints)
case !reflect.DeepEqual(expected.OCSPServers, cert.OCSPServer):
return fmt.Errorf("OCSPServer:\nexpected\n%#v\ngot\n%#v\n", expected.OCSPServers, cert.OCSPServer)
2018-02-10 15:07:10 +00:00
case !reflect.DeepEqual([]string{"intermediate.cert.com"}, cert.DNSNames):
return fmt.Errorf("DNSNames\nexpected\n%#v\ngot\n%#v\n", []string{"intermediate.cert.com"}, cert.DNSNames)
case !reflect.DeepEqual(x509.SHA512WithRSA, cert.SignatureAlgorithm):
return fmt.Errorf("Signature Algorithm:\nexpected\n%#v\ngot\n%#v\n", x509.SHA512WithRSA, cert.SignatureAlgorithm)
case !reflect.DeepEqual(skid, cert.SubjectKeyId):
return fmt.Errorf("SKID:\nexpected\n%#v\ngot\n%#v\n", skid, cert.SubjectKeyId)
}
if math.Abs(float64(time.Now().Add(-2*time.Hour).Unix()-cert.NotBefore.Unix())) > 10 {
t.Fatalf("root/sign-intermediate did not properly set validity period (notBefore): was %v vs expected %v", cert.NotBefore, time.Now().Add(-2*time.Hour))
}
return nil
},
},
// Same as above but exclude adding to sans
{
Operation: logical.UpdateOperation,
Path: "root/sign-intermediate",
Data: map[string]interface{}{
2018-02-10 15:07:10 +00:00
"common_name": "intermediate.cert.com",
"csr": csrPem2048,
"format": "der",
"exclude_cn_from_sans": true,
},
Check: func(resp *logical.Response) error {
certString := resp.Data["certificate"].(string)
if certString == "" {
return fmt.Errorf("no certificate returned")
}
if resp.Secret != nil && resp.Secret.LeaseID != "" {
return fmt.Errorf("signed intermediate returned with a lease")
}
certBytes, _ := base64.StdEncoding.DecodeString(certString)
certs, err := x509.ParseCertificates(certBytes)
if err != nil {
return fmt.Errorf("returned cert cannot be parsed: %w", err)
}
if len(certs) != 1 {
return fmt.Errorf("unexpected returned length of certificates: %d", len(certs))
}
cert := certs[0]
switch {
case !reflect.DeepEqual(expected.IssuingCertificates, cert.IssuingCertificateURL):
return fmt.Errorf("expected\n%#v\ngot\n%#v\n", expected.IssuingCertificates, cert.IssuingCertificateURL)
case !reflect.DeepEqual(expected.CRLDistributionPoints, cert.CRLDistributionPoints):
return fmt.Errorf("expected\n%#v\ngot\n%#v\n", expected.CRLDistributionPoints, cert.CRLDistributionPoints)
case !reflect.DeepEqual(expected.OCSPServers, cert.OCSPServer):
return fmt.Errorf("expected\n%#v\ngot\n%#v\n", expected.OCSPServers, cert.OCSPServer)
case !reflect.DeepEqual([]string(nil), cert.DNSNames):
return fmt.Errorf("expected\n%#v\ngot\n%#v\n", []string(nil), cert.DNSNames)
}
return nil
},
},
}
return ret
}
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
func generateCSR(t *testing.T, csrTemplate *x509.CertificateRequest, keyType string, keyBits int) (interface{}, []byte, string) {
var priv interface{}
var err error
switch keyType {
case "rsa":
priv, err = rsa.GenerateKey(rand.Reader, keyBits)
case "ec":
switch keyBits {
case 224:
priv, err = ecdsa.GenerateKey(elliptic.P224(), rand.Reader)
case 256:
priv, err = ecdsa.GenerateKey(elliptic.P256(), rand.Reader)
case 384:
priv, err = ecdsa.GenerateKey(elliptic.P384(), rand.Reader)
case 521:
priv, err = ecdsa.GenerateKey(elliptic.P521(), rand.Reader)
default:
t.Fatalf("Got unknown ec< key bits: %v", keyBits)
}
case "ed25519":
_, priv, err = ed25519.GenerateKey(rand.Reader)
}
if err != nil {
t.Fatalf("Got error generating private key for CSR: %v", err)
}
csr, err := x509.CreateCertificateRequest(rand.Reader, csrTemplate, priv)
if err != nil {
t.Fatalf("Got error generating CSR: %v", err)
}
csrPem := strings.TrimSpace(string(pem.EncodeToMemory(&pem.Block{
Type: "CERTIFICATE REQUEST",
Bytes: csr,
})))
return priv, csr, csrPem
}
2015-10-14 15:46:01 +00:00
func generateCSRSteps(t *testing.T, caCert, caKey string, intdata, reqdata map[string]interface{}) []logicaltest.TestStep {
csrTemplate, csrPem := generateTestCsr(t, certutil.RSAPrivateKey, 2048)
2015-10-14 15:46:01 +00:00
ret := []logicaltest.TestStep{
{
2016-01-07 15:30:47 +00:00
Operation: logical.UpdateOperation,
Path: "root/generate/exported",
2015-10-14 15:46:01 +00:00
Data: map[string]interface{}{
"common_name": "Root Cert",
"ttl": "180h",
"max_path_length": 0,
},
},
{
2016-01-07 15:30:47 +00:00
Operation: logical.UpdateOperation,
Path: "root/sign-intermediate",
Data: map[string]interface{}{
"use_csr_values": true,
"csr": csrPem,
"format": "der",
},
ErrorOk: true,
},
{
Operation: logical.DeleteOperation,
Path: "root",
},
{
2016-01-07 15:30:47 +00:00
Operation: logical.UpdateOperation,
Path: "root/generate/exported",
Data: map[string]interface{}{
"common_name": "Root Cert",
"ttl": "180h",
"max_path_length": 1,
2015-10-14 15:46:01 +00:00
},
},
{
2016-01-07 15:30:47 +00:00
Operation: logical.UpdateOperation,
Path: "root/sign-intermediate",
2015-10-14 15:46:01 +00:00
Data: map[string]interface{}{
"use_csr_values": true,
"csr": csrPem,
2015-10-14 15:46:01 +00:00
"format": "der",
},
Check: func(resp *logical.Response) error {
certString := resp.Data["certificate"].(string)
if certString == "" {
return fmt.Errorf("no certificate returned")
}
certBytes, _ := base64.StdEncoding.DecodeString(certString)
certs, err := x509.ParseCertificates(certBytes)
if err != nil {
return fmt.Errorf("returned cert cannot be parsed: %w", err)
2015-10-14 15:46:01 +00:00
}
if len(certs) != 1 {
return fmt.Errorf("unexpected returned length of certificates: %d", len(certs))
}
cert := certs[0]
if cert.MaxPathLen != 0 {
return fmt.Errorf("max path length of %d does not match the requested of 3", cert.MaxPathLen)
}
if !cert.MaxPathLenZero {
return fmt.Errorf("max path length zero is not set")
}
2015-10-14 15:46:01 +00:00
// We need to set these as they are filled in with unparsed values in the final cert
csrTemplate.Subject.Names = cert.Subject.Names
csrTemplate.Subject.ExtraNames = cert.Subject.ExtraNames
2015-10-14 15:46:01 +00:00
switch {
case !reflect.DeepEqual(cert.Subject, csrTemplate.Subject):
return fmt.Errorf("cert subject\n%#v\ndoes not match csr subject\n%#v\n", cert.Subject, csrTemplate.Subject)
case !reflect.DeepEqual(cert.DNSNames, csrTemplate.DNSNames):
return fmt.Errorf("cert dns names\n%#v\ndoes not match csr dns names\n%#v\n", cert.DNSNames, csrTemplate.DNSNames)
case !reflect.DeepEqual(cert.EmailAddresses, csrTemplate.EmailAddresses):
return fmt.Errorf("cert email addresses\n%#v\ndoes not match csr email addresses\n%#v\n", cert.EmailAddresses, csrTemplate.EmailAddresses)
case !reflect.DeepEqual(cert.IPAddresses, csrTemplate.IPAddresses):
return fmt.Errorf("cert ip addresses\n%#v\ndoes not match csr ip addresses\n%#v\n", cert.IPAddresses, csrTemplate.IPAddresses)
}
return nil
},
},
}
return ret
}
func generateTestCsr(t *testing.T, keyType certutil.PrivateKeyType, keyBits int) (x509.CertificateRequest, string) {
csrTemplate := x509.CertificateRequest{
Subject: pkix.Name{
Country: []string{"MyCountry"},
PostalCode: []string{"MyPostalCode"},
SerialNumber: "MySerialNumber",
CommonName: "my@example.com",
},
DNSNames: []string{
"name1.example.com",
"name2.example.com",
"name3.example.com",
},
EmailAddresses: []string{
"name1@example.com",
"name2@example.com",
"name3@example.com",
},
IPAddresses: []net.IP{
net.ParseIP("::ff:1:2:3:4"),
net.ParseIP("::ff:5:6:7:8"),
},
}
_, _, csrPem := generateCSR(t, &csrTemplate, string(keyType), keyBits)
return csrTemplate, csrPem
}
// Generates steps to test out various role permutations
func generateRoleSteps(t *testing.T, useCSRs bool) []logicaltest.TestStep {
roleVals := roleEntry{
Add role parameter to restrict issuance of wildcard certificates (#14238) * Add new AllowWildcardCertificate field to PKI role This field allows the PKI role to control whether or not issuance of wildcard certificates are allowed. We default (both on migration and new role creation) to the less secure true value for backwards compatibility with existing Vault versions. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sanitizedName to reducedName Per comment, this variable name was confusing during the reproduction and subsequent fix of the earlier vulnerability and associated bug report. Because the common name isn't necessarily _sanitized_ in any way (and indeed must be considered in relation to other parts or the whole), but portions of the entire name are removed, reducedName appears to make the most sense. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Enforce AllowWildcardCertificates during issuance This commit adds the bulk of correctly validating wildcard certificate Common Names during issuance according to RFC 6125 Section 6.4.3 semantics. As part of this, support for RFC 2818-conforming wildcard certificates (wherein there are almost no restrictions on issuance) has been removed. Note that this flag does take precedence over AllowAnyName, giving a little more safety in wildcard issuance in this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test cases to conform with RFC 6125 Test cases 19, 70+71, and 83+84 didn't conform with the RFC 6125, and so should've been rejected under strict conformance. For 70+71 and 83+84, we previously conditioned around the value of AllowSubdomains (allowing issuance when true), but they likely should've been rejected either way. Additionally, update the notes about globs matching wildcard certificates to notate this is indeed the case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Check AllowWildcardCertifciates in issuance tests This allows for regression tests to cover the new AllowWildcardCertificate conditional. We add additional test cases ensuring that wildcard issuance is properly forbidden in all relevant scenarios, while allowing the existing test cases to validate that wildcard status doesn't affect non-wildcard certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add Wildcard allowance during signing operations When using sign-verbatim, sign-intermediate, or getting certificate generation parameters, set AllowWildcardCertificates to mirror existing policies. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-02-24 13:41:56 +00:00
MaxTTL: 12 * time.Hour,
KeyType: "rsa",
KeyBits: 2048,
RequireCN: true,
AllowWildcardCertificates: new(bool),
}
Add role parameter to restrict issuance of wildcard certificates (#14238) * Add new AllowWildcardCertificate field to PKI role This field allows the PKI role to control whether or not issuance of wildcard certificates are allowed. We default (both on migration and new role creation) to the less secure true value for backwards compatibility with existing Vault versions. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sanitizedName to reducedName Per comment, this variable name was confusing during the reproduction and subsequent fix of the earlier vulnerability and associated bug report. Because the common name isn't necessarily _sanitized_ in any way (and indeed must be considered in relation to other parts or the whole), but portions of the entire name are removed, reducedName appears to make the most sense. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Enforce AllowWildcardCertificates during issuance This commit adds the bulk of correctly validating wildcard certificate Common Names during issuance according to RFC 6125 Section 6.4.3 semantics. As part of this, support for RFC 2818-conforming wildcard certificates (wherein there are almost no restrictions on issuance) has been removed. Note that this flag does take precedence over AllowAnyName, giving a little more safety in wildcard issuance in this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test cases to conform with RFC 6125 Test cases 19, 70+71, and 83+84 didn't conform with the RFC 6125, and so should've been rejected under strict conformance. For 70+71 and 83+84, we previously conditioned around the value of AllowSubdomains (allowing issuance when true), but they likely should've been rejected either way. Additionally, update the notes about globs matching wildcard certificates to notate this is indeed the case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Check AllowWildcardCertifciates in issuance tests This allows for regression tests to cover the new AllowWildcardCertificate conditional. We add additional test cases ensuring that wildcard issuance is properly forbidden in all relevant scenarios, while allowing the existing test cases to validate that wildcard status doesn't affect non-wildcard certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add Wildcard allowance during signing operations When using sign-verbatim, sign-intermediate, or getting certificate generation parameters, set AllowWildcardCertificates to mirror existing policies. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-02-24 13:41:56 +00:00
*roleVals.AllowWildcardCertificates = true
issueVals := certutil.IssueData{}
ret := []logicaltest.TestStep{}
roleTestStep := logicaltest.TestStep{
2016-01-07 15:30:47 +00:00
Operation: logical.UpdateOperation,
Path: "roles/test",
}
var issueTestStep logicaltest.TestStep
if useCSRs {
issueTestStep = logicaltest.TestStep{
2016-01-07 15:30:47 +00:00
Operation: logical.UpdateOperation,
Path: "sign/test",
}
} else {
issueTestStep = logicaltest.TestStep{
2016-01-07 15:30:47 +00:00
Operation: logical.UpdateOperation,
Path: "issue/test",
}
}
generatedRSAKeys := map[int]crypto.Signer{}
generatedECKeys := map[int]crypto.Signer{}
generatedEdKeys := map[int]crypto.Signer{}
/*
// For the number of tests being run, a seed of 1 has been tested
// to hit all of the various values below. However, for normal
// testing we use a randomized time for maximum fuzziness.
*/
var seed int64 = 1
fixedSeed := os.Getenv("VAULT_PKITESTS_FIXED_SEED")
if len(fixedSeed) == 0 {
seed = time.Now().UnixNano()
} else {
var err error
seed, err = strconv.ParseInt(fixedSeed, 10, 64)
if err != nil {
t.Fatalf("error parsing fixed seed of %s: %v", fixedSeed, err)
}
}
mathRand := mathrand.New(mathrand.NewSource(seed))
// t.Logf("seed under test: %v", seed)
// Used by tests not toggling common names to turn off the behavior of random key bit fuzziness
keybitSizeRandOff := false
genericErrorOkCheck := func(resp *logical.Response) error {
if resp.IsError() {
return nil
}
return fmt.Errorf("expected an error, but did not seem to get one")
}
// Adds tests with the currently configured issue/role information
addTests := func(testCheck logicaltest.TestCheckFunc) {
2016-07-19 17:54:18 +00:00
stepCount++
// t.Logf("test step %d\nrole vals: %#v\n", stepCount, roleVals)
2016-07-19 17:54:18 +00:00
stepCount++
// t.Logf("test step %d\nissue vals: %#v\n", stepCount, issueTestStep)
roleTestStep.Data = roleVals.ToResponseData()
roleTestStep.Data["generate_lease"] = false
ret = append(ret, roleTestStep)
issueTestStep.Data = structs.New(issueVals).Map()
switch {
case issueTestStep.ErrorOk:
issueTestStep.Check = genericErrorOkCheck
case testCheck != nil:
issueTestStep.Check = testCheck
default:
issueTestStep.Check = nil
}
ret = append(ret, issueTestStep)
}
getCountryCheck := func(role roleEntry) logicaltest.TestCheckFunc {
var certBundle certutil.CertBundle
return func(resp *logical.Response) error {
err := mapstructure.Decode(resp.Data, &certBundle)
if err != nil {
return err
}
parsedCertBundle, err := certBundle.ToParsedCertBundle()
if err != nil {
return fmt.Errorf("error checking generated certificate: %s", err)
}
cert := parsedCertBundle.Certificate
expected := strutil.RemoveDuplicates(role.Country, true)
if !reflect.DeepEqual(cert.Subject.Country, expected) {
return fmt.Errorf("error: returned certificate has Country of %s but %s was specified in the role", cert.Subject.Country, expected)
}
return nil
}
}
getOuCheck := func(role roleEntry) logicaltest.TestCheckFunc {
var certBundle certutil.CertBundle
return func(resp *logical.Response) error {
err := mapstructure.Decode(resp.Data, &certBundle)
if err != nil {
return err
}
parsedCertBundle, err := certBundle.ToParsedCertBundle()
if err != nil {
return fmt.Errorf("error checking generated certificate: %s", err)
}
cert := parsedCertBundle.Certificate
expected := strutil.RemoveDuplicatesStable(role.OU, true)
if !reflect.DeepEqual(cert.Subject.OrganizationalUnit, expected) {
return fmt.Errorf("error: returned certificate has OU of %s but %s was specified in the role", cert.Subject.OrganizationalUnit, expected)
}
return nil
}
}
getOrganizationCheck := func(role roleEntry) logicaltest.TestCheckFunc {
var certBundle certutil.CertBundle
return func(resp *logical.Response) error {
err := mapstructure.Decode(resp.Data, &certBundle)
if err != nil {
return err
}
parsedCertBundle, err := certBundle.ToParsedCertBundle()
if err != nil {
return fmt.Errorf("error checking generated certificate: %s", err)
}
cert := parsedCertBundle.Certificate
expected := strutil.RemoveDuplicates(role.Organization, true)
if !reflect.DeepEqual(cert.Subject.Organization, expected) {
return fmt.Errorf("error: returned certificate has Organization of %s but %s was specified in the role", cert.Subject.Organization, expected)
}
return nil
}
}
getLocalityCheck := func(role roleEntry) logicaltest.TestCheckFunc {
var certBundle certutil.CertBundle
return func(resp *logical.Response) error {
err := mapstructure.Decode(resp.Data, &certBundle)
if err != nil {
return err
}
parsedCertBundle, err := certBundle.ToParsedCertBundle()
if err != nil {
return fmt.Errorf("error checking generated certificate: %s", err)
}
cert := parsedCertBundle.Certificate
expected := strutil.RemoveDuplicates(role.Locality, true)
if !reflect.DeepEqual(cert.Subject.Locality, expected) {
return fmt.Errorf("error: returned certificate has Locality of %s but %s was specified in the role", cert.Subject.Locality, expected)
}
return nil
}
}
getProvinceCheck := func(role roleEntry) logicaltest.TestCheckFunc {
var certBundle certutil.CertBundle
return func(resp *logical.Response) error {
err := mapstructure.Decode(resp.Data, &certBundle)
if err != nil {
return err
}
parsedCertBundle, err := certBundle.ToParsedCertBundle()
if err != nil {
return fmt.Errorf("error checking generated certificate: %s", err)
}
cert := parsedCertBundle.Certificate
expected := strutil.RemoveDuplicates(role.Province, true)
if !reflect.DeepEqual(cert.Subject.Province, expected) {
return fmt.Errorf("error: returned certificate has Province of %s but %s was specified in the role", cert.Subject.Province, expected)
}
return nil
}
}
getStreetAddressCheck := func(role roleEntry) logicaltest.TestCheckFunc {
var certBundle certutil.CertBundle
return func(resp *logical.Response) error {
err := mapstructure.Decode(resp.Data, &certBundle)
if err != nil {
return err
}
parsedCertBundle, err := certBundle.ToParsedCertBundle()
if err != nil {
return fmt.Errorf("error checking generated certificate: %s", err)
}
cert := parsedCertBundle.Certificate
expected := strutil.RemoveDuplicates(role.StreetAddress, true)
if !reflect.DeepEqual(cert.Subject.StreetAddress, expected) {
return fmt.Errorf("error: returned certificate has StreetAddress of %s but %s was specified in the role", cert.Subject.StreetAddress, expected)
}
return nil
}
}
getPostalCodeCheck := func(role roleEntry) logicaltest.TestCheckFunc {
var certBundle certutil.CertBundle
return func(resp *logical.Response) error {
err := mapstructure.Decode(resp.Data, &certBundle)
if err != nil {
return err
}
parsedCertBundle, err := certBundle.ToParsedCertBundle()
if err != nil {
return fmt.Errorf("error checking generated certificate: %s", err)
}
cert := parsedCertBundle.Certificate
expected := strutil.RemoveDuplicates(role.PostalCode, true)
if !reflect.DeepEqual(cert.Subject.PostalCode, expected) {
return fmt.Errorf("error: returned certificate has PostalCode of %s but %s was specified in the role", cert.Subject.PostalCode, expected)
}
return nil
}
}
getNotBeforeCheck := func(role roleEntry) logicaltest.TestCheckFunc {
var certBundle certutil.CertBundle
return func(resp *logical.Response) error {
err := mapstructure.Decode(resp.Data, &certBundle)
if err != nil {
return err
}
parsedCertBundle, err := certBundle.ToParsedCertBundle()
if err != nil {
return fmt.Errorf("error checking generated certificate: %s", err)
}
cert := parsedCertBundle.Certificate
actualDiff := time.Since(cert.NotBefore)
certRoleDiff := (role.NotBeforeDuration - actualDiff).Truncate(time.Second)
// These times get truncated, so give a 1 second buffer on each side
if certRoleDiff >= -1*time.Second && certRoleDiff <= 1*time.Second {
return nil
}
return fmt.Errorf("validity period out of range diff: %v", certRoleDiff)
}
}
// Returns a TestCheckFunc that performs various validity checks on the
// returned certificate information, mostly within checkCertsAndPrivateKey
getCnCheck := func(name string, role roleEntry, key crypto.Signer, usage x509.KeyUsage, extUsage x509.ExtKeyUsage, validity time.Duration) logicaltest.TestCheckFunc {
var certBundle certutil.CertBundle
return func(resp *logical.Response) error {
err := mapstructure.Decode(resp.Data, &certBundle)
if err != nil {
return err
}
parsedCertBundle, err := checkCertsAndPrivateKey(role.KeyType, key, usage, extUsage, validity, &certBundle)
if err != nil {
return fmt.Errorf("error checking generated certificate: %s", err)
}
cert := parsedCertBundle.Certificate
if cert.Subject.CommonName != name {
return fmt.Errorf("error: returned certificate has CN of %s but %s was requested", cert.Subject.CommonName, name)
}
if strings.Contains(cert.Subject.CommonName, "@") {
if len(cert.DNSNames) != 0 || len(cert.EmailAddresses) != 1 {
return fmt.Errorf("error: found more than one DNS SAN or not one Email SAN but only one was requested, cert.DNSNames = %#v, cert.EmailAddresses = %#v", cert.DNSNames, cert.EmailAddresses)
}
} else {
if len(cert.DNSNames) != 1 || len(cert.EmailAddresses) != 0 {
return fmt.Errorf("error: found more than one Email SAN or not one DNS SAN but only one was requested, cert.DNSNames = %#v, cert.EmailAddresses = %#v", cert.DNSNames, cert.EmailAddresses)
}
}
var retName string
if len(cert.DNSNames) > 0 {
retName = cert.DNSNames[0]
}
if len(cert.EmailAddresses) > 0 {
retName = cert.EmailAddresses[0]
}
if retName != name {
2018-02-10 15:07:10 +00:00
// Check IDNA
p := idna.New(
idna.StrictDomainName(true),
idna.VerifyDNSLength(true),
)
converted, err := p.ToUnicode(retName)
if err != nil {
t.Fatal(err)
}
if converted != name {
return fmt.Errorf("error: returned certificate has a DNS SAN of %s (from idna: %s) but %s was requested", retName, converted, name)
2018-02-10 15:07:10 +00:00
}
}
return nil
}
}
type csrPlan struct {
errorOk bool
roleKeyBits int
cert string
privKey crypto.Signer
}
getCsr := func(keyType string, keyBits int, csrTemplate *x509.CertificateRequest) (*pem.Block, crypto.Signer) {
var privKey crypto.Signer
var ok bool
switch keyType {
case "rsa":
privKey, ok = generatedRSAKeys[keyBits]
if !ok {
privKey, _ = rsa.GenerateKey(rand.Reader, keyBits)
generatedRSAKeys[keyBits] = privKey
}
case "ec":
var curve elliptic.Curve
switch keyBits {
case 224:
curve = elliptic.P224()
case 256:
curve = elliptic.P256()
case 384:
curve = elliptic.P384()
case 521:
curve = elliptic.P521()
}
privKey, ok = generatedECKeys[keyBits]
if !ok {
privKey, _ = ecdsa.GenerateKey(curve, rand.Reader)
generatedECKeys[keyBits] = privKey
}
case "ed25519":
privKey, ok = generatedEdKeys[keyBits]
if !ok {
_, privKey, _ = ed25519.GenerateKey(rand.Reader)
generatedEdKeys[keyBits] = privKey
}
default:
panic("invalid key type: " + keyType)
}
csr, err := x509.CreateCertificateRequest(rand.Reader, csrTemplate, privKey)
if err != nil {
t.Fatalf("Error creating certificate request: %s", err)
}
block := pem.Block{
Type: "CERTIFICATE REQUEST",
Bytes: csr,
}
return &block, privKey
}
getRandCsr := func(keyType string, errorOk bool, csrTemplate *x509.CertificateRequest) csrPlan {
rsaKeyBits := []int{2048, 3072, 4096}
ecKeyBits := []int{224, 256, 384, 521}
plan := csrPlan{errorOk: errorOk}
var testBitSize int
switch keyType {
case "rsa":
plan.roleKeyBits = rsaKeyBits[mathRand.Int()%len(rsaKeyBits)]
testBitSize = plan.roleKeyBits
// If we don't expect an error already, randomly choose a
// key size and expect an error if it's less than the role
// setting
if !keybitSizeRandOff && !errorOk {
testBitSize = rsaKeyBits[mathRand.Int()%len(rsaKeyBits)]
}
if testBitSize < plan.roleKeyBits {
plan.errorOk = true
}
case "ec":
plan.roleKeyBits = ecKeyBits[mathRand.Int()%len(ecKeyBits)]
testBitSize = plan.roleKeyBits
// If we don't expect an error already, randomly choose a
// key size and expect an error if it's less than the role
// setting
if !keybitSizeRandOff && !errorOk {
testBitSize = ecKeyBits[mathRand.Int()%len(ecKeyBits)]
}
if testBitSize < plan.roleKeyBits {
plan.errorOk = true
}
default:
panic("invalid key type: " + keyType)
}
if len(os.Getenv("VAULT_VERBOSE_PKITESTS")) > 0 {
t.Logf("roleKeyBits=%d testBitSize=%d errorOk=%v", plan.roleKeyBits, testBitSize, plan.errorOk)
}
block, privKey := getCsr(keyType, testBitSize, csrTemplate)
plan.cert = strings.TrimSpace(string(pem.EncodeToMemory(block)))
plan.privKey = privKey
return plan
}
// Common names to test with the various role flags toggled
var commonNames struct {
Localhost bool `structs:"localhost"`
BareDomain bool `structs:"example.com"`
SecondDomain bool `structs:"foobar.com"`
SubDomain bool `structs:"foo.example.com"`
Wildcard bool `structs:"*.example.com"`
SubSubdomain bool `structs:"foo.bar.example.com"`
SubSubdomainWildcard bool `structs:"*.bar.example.com"`
GlobDomain bool `structs:"fooexample.com"`
2018-02-10 15:07:10 +00:00
IDN bool `structs:"daɪˈɛrɨsɨs"`
AnyHost bool `structs:"porkslap.beer"`
}
// Adds a series of tests based on the current selection of
// allowed common names; contains some (seeded) randomness
//
// This allows for a variety of common names to be tested in various
// combinations with allowed toggles of the role
addCnTests := func() {
cnMap := structs.New(commonNames).Map()
for name, allowedInt := range cnMap {
roleVals.KeyType = "rsa"
roleVals.KeyBits = 2048
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
if mathRand.Int()%3 == 1 {
roleVals.KeyType = "ec"
roleVals.KeyBits = 224
}
roleVals.ServerFlag = false
roleVals.ClientFlag = false
roleVals.CodeSigningFlag = false
roleVals.EmailProtectionFlag = false
var usage []string
if mathRand.Int()%2 == 1 {
usage = append(usage, "DigitalSignature")
}
if mathRand.Int()%2 == 1 {
usage = append(usage, "ContentCoMmitment")
}
if mathRand.Int()%2 == 1 {
usage = append(usage, "KeyEncipherment")
}
if mathRand.Int()%2 == 1 {
usage = append(usage, "DataEncipherment")
}
if mathRand.Int()%2 == 1 {
usage = append(usage, "KeyAgreemEnt")
}
if mathRand.Int()%2 == 1 {
usage = append(usage, "CertSign")
}
if mathRand.Int()%2 == 1 {
usage = append(usage, "CRLSign")
}
if mathRand.Int()%2 == 1 {
usage = append(usage, "EncipherOnly")
}
if mathRand.Int()%2 == 1 {
usage = append(usage, "DecipherOnly")
}
roleVals.KeyUsage = usage
parsedKeyUsage := parseKeyUsages(roleVals.KeyUsage)
if parsedKeyUsage == 0 && len(usage) != 0 {
2016-06-23 13:49:03 +00:00
panic("parsed key usages was zero")
}
var extUsage x509.ExtKeyUsage
i := mathRand.Int() % 4
switch {
case i == 0:
2018-02-20 05:03:45 +00:00
// Punt on this for now since I'm not clear the actual proper
// way to format these
if name != "daɪˈɛrɨsɨs" {
extUsage = x509.ExtKeyUsageEmailProtection
roleVals.EmailProtectionFlag = true
break
}
fallthrough
case i == 1:
extUsage = x509.ExtKeyUsageServerAuth
roleVals.ServerFlag = true
case i == 2:
extUsage = x509.ExtKeyUsageClientAuth
roleVals.ClientFlag = true
default:
extUsage = x509.ExtKeyUsageCodeSigning
roleVals.CodeSigningFlag = true
}
allowed := allowedInt.(bool)
issueVals.CommonName = name
if roleVals.EmailProtectionFlag {
if !strings.HasPrefix(name, "*") {
issueVals.CommonName = "user@" + issueVals.CommonName
}
}
issueTestStep.ErrorOk = !allowed
validity := roleVals.MaxTTL
if useCSRs {
templ := &x509.CertificateRequest{
Subject: pkix.Name{
CommonName: issueVals.CommonName,
},
}
plan := getRandCsr(roleVals.KeyType, issueTestStep.ErrorOk, templ)
issueVals.CSR = plan.cert
roleVals.KeyBits = plan.roleKeyBits
issueTestStep.ErrorOk = plan.errorOk
addTests(getCnCheck(issueVals.CommonName, roleVals, plan.privKey, x509.KeyUsage(parsedKeyUsage), extUsage, validity))
} else {
addTests(getCnCheck(issueVals.CommonName, roleVals, nil, x509.KeyUsage(parsedKeyUsage), extUsage, validity))
}
}
}
funcs := []interface{}{
addCnTests, getCnCheck, getCountryCheck, getLocalityCheck, getNotBeforeCheck,
getOrganizationCheck, getOuCheck, getPostalCodeCheck, getRandCsr, getStreetAddressCheck,
getProvinceCheck,
}
if len(os.Getenv("VAULT_VERBOSE_PKITESTS")) > 0 {
t.Logf("funcs=%d", len(funcs))
}
// Common Name tests
{
// common_name not provided
issueVals.CommonName = ""
issueTestStep.ErrorOk = true
addTests(nil)
// Nothing is allowed
addCnTests()
roleVals.AllowLocalhost = true
commonNames.Localhost = true
addCnTests()
roleVals.AllowedDomains = []string{"foobar.com"}
addCnTests()
roleVals.AllowedDomains = []string{"example.com"}
roleVals.AllowSubdomains = true
commonNames.SubDomain = true
commonNames.Wildcard = true
commonNames.SubSubdomain = true
commonNames.SubSubdomainWildcard = true
addCnTests()
roleVals.AllowedDomains = []string{"foobar.com", "example.com"}
commonNames.SecondDomain = true
roleVals.AllowBareDomains = true
commonNames.BareDomain = true
addCnTests()
roleVals.AllowedDomains = []string{"foobar.com", "*example.com"}
roleVals.AllowGlobDomains = true
commonNames.GlobDomain = true
addCnTests()
roleVals.AllowAnyName = true
roleVals.EnforceHostnames = true
commonNames.AnyHost = true
2018-02-10 15:07:10 +00:00
commonNames.IDN = true
addCnTests()
roleVals.EnforceHostnames = false
addCnTests()
// Ensure that we end up with acceptable key sizes since they won't be
// toggled any longer
keybitSizeRandOff = true
addCnTests()
}
// Country tests
{
roleVals.Country = []string{"foo"}
addTests(getCountryCheck(roleVals))
roleVals.Country = []string{"foo", "bar"}
addTests(getCountryCheck(roleVals))
}
// OU tests
{
roleVals.OU = []string{"foo"}
addTests(getOuCheck(roleVals))
roleVals.OU = []string{"bar", "foo"}
addTests(getOuCheck(roleVals))
}
// Organization tests
{
roleVals.Organization = []string{"system:masters"}
addTests(getOrganizationCheck(roleVals))
roleVals.Organization = []string{"foo", "bar"}
addTests(getOrganizationCheck(roleVals))
}
// Locality tests
{
roleVals.Locality = []string{"foo"}
addTests(getLocalityCheck(roleVals))
roleVals.Locality = []string{"foo", "bar"}
addTests(getLocalityCheck(roleVals))
}
// Province tests
{
roleVals.Province = []string{"foo"}
addTests(getProvinceCheck(roleVals))
roleVals.Province = []string{"foo", "bar"}
addTests(getProvinceCheck(roleVals))
}
// StreetAddress tests
{
roleVals.StreetAddress = []string{"123 foo street"}
addTests(getStreetAddressCheck(roleVals))
roleVals.StreetAddress = []string{"123 foo street", "456 bar avenue"}
addTests(getStreetAddressCheck(roleVals))
}
// PostalCode tests
{
roleVals.PostalCode = []string{"f00"}
addTests(getPostalCodeCheck(roleVals))
roleVals.PostalCode = []string{"f00", "b4r"}
addTests(getPostalCodeCheck(roleVals))
}
// NotBefore tests
{
roleVals.NotBeforeDuration = 10 * time.Second
addTests(getNotBeforeCheck(roleVals))
roleVals.NotBeforeDuration = 30 * time.Second
addTests(getNotBeforeCheck(roleVals))
roleVals.NotBeforeDuration = 0
}
// IP SAN tests
{
getIpCheck := func(expectedIp ...net.IP) logicaltest.TestCheckFunc {
return func(resp *logical.Response) error {
var certBundle certutil.CertBundle
err := mapstructure.Decode(resp.Data, &certBundle)
if err != nil {
return err
}
parsedCertBundle, err := certBundle.ToParsedCertBundle()
if err != nil {
return fmt.Errorf("error parsing cert bundle: %s", err)
}
cert := parsedCertBundle.Certificate
var expected []net.IP
expected = append(expected, expectedIp...)
if diff := deep.Equal(cert.IPAddresses, expected); len(diff) > 0 {
return fmt.Errorf("wrong SAN IPs, diff: %v", diff)
}
return nil
}
}
addIPSANTests := func(useCSRs, useCSRSANs, allowIPSANs, errorOk bool, ipSANs string, csrIPSANs []net.IP, check logicaltest.TestCheckFunc) {
if useCSRs {
csrTemplate := &x509.CertificateRequest{
Subject: pkix.Name{
CommonName: issueVals.CommonName,
},
IPAddresses: csrIPSANs,
}
block, _ := getCsr(roleVals.KeyType, roleVals.KeyBits, csrTemplate)
issueVals.CSR = strings.TrimSpace(string(pem.EncodeToMemory(block)))
}
oldRoleVals, oldIssueVals, oldIssueTestStep := roleVals, issueVals, issueTestStep
roleVals.UseCSRSANs = useCSRSANs
roleVals.AllowIPSANs = allowIPSANs
issueVals.CommonName = "someone@example.com"
issueVals.IPSANs = ipSANs
issueTestStep.ErrorOk = errorOk
addTests(check)
roleVals, issueVals, issueTestStep = oldRoleVals, oldIssueVals, oldIssueTestStep
}
roleVals.AllowAnyName = true
roleVals.EnforceHostnames = true
roleVals.AllowLocalhost = true
roleVals.UseCSRCommonName = true
commonNames.Localhost = true
netip1, netip2 := net.IP{127, 0, 0, 1}, net.IP{170, 171, 172, 173}
textip1, textip3 := "127.0.0.1", "::1"
// IPSANs not allowed and not provided, should not be an error.
addIPSANTests(useCSRs, false, false, false, "", nil, getIpCheck())
// IPSANs not allowed, valid IPSANs provided, should be an error.
addIPSANTests(useCSRs, false, false, true, textip1+","+textip3, nil, nil)
// IPSANs allowed, bogus IPSANs provided, should be an error.
addIPSANTests(useCSRs, false, true, true, "foobar", nil, nil)
// Given IPSANs as API argument and useCSRSANs false, CSR arg ignored.
addIPSANTests(useCSRs, false, true, false, textip1,
[]net.IP{netip2}, getIpCheck(netip1))
if useCSRs {
// IPSANs not allowed, valid IPSANs provided via CSR, should be an error.
addIPSANTests(useCSRs, true, false, true, "", []net.IP{netip1}, nil)
// Given IPSANs as both API and CSR arguments and useCSRSANs=true, API arg ignored.
addIPSANTests(useCSRs, true, true, false, textip3,
[]net.IP{netip1, netip2}, getIpCheck(netip1, netip2))
}
}
{
getOtherCheck := func(expectedOthers ...otherNameUtf8) logicaltest.TestCheckFunc {
return func(resp *logical.Response) error {
var certBundle certutil.CertBundle
err := mapstructure.Decode(resp.Data, &certBundle)
if err != nil {
return err
}
parsedCertBundle, err := certBundle.ToParsedCertBundle()
if err != nil {
return fmt.Errorf("error parsing cert bundle: %s", err)
}
cert := parsedCertBundle.Certificate
foundOthers, err := getOtherSANsFromX509Extensions(cert.Extensions)
if err != nil {
return err
}
var expected []otherNameUtf8
expected = append(expected, expectedOthers...)
if diff := deep.Equal(foundOthers, expected); len(diff) > 0 {
return fmt.Errorf("wrong SAN IPs, diff: %v", diff)
}
return nil
}
}
addOtherSANTests := func(useCSRs, useCSRSANs bool, allowedOtherSANs []string, errorOk bool, otherSANs []string, csrOtherSANs []otherNameUtf8, check logicaltest.TestCheckFunc) {
otherSansMap := func(os []otherNameUtf8) map[string][]string {
ret := make(map[string][]string)
for _, o := range os {
ret[o.oid] = append(ret[o.oid], o.value)
}
return ret
}
if useCSRs {
csrTemplate := &x509.CertificateRequest{
Subject: pkix.Name{
CommonName: issueVals.CommonName,
},
}
if err := handleOtherCSRSANs(csrTemplate, otherSansMap(csrOtherSANs)); err != nil {
t.Fatal(err)
}
block, _ := getCsr(roleVals.KeyType, roleVals.KeyBits, csrTemplate)
issueVals.CSR = strings.TrimSpace(string(pem.EncodeToMemory(block)))
}
oldRoleVals, oldIssueVals, oldIssueTestStep := roleVals, issueVals, issueTestStep
roleVals.UseCSRSANs = useCSRSANs
roleVals.AllowedOtherSANs = allowedOtherSANs
issueVals.CommonName = "someone@example.com"
issueVals.OtherSANs = strings.Join(otherSANs, ",")
issueTestStep.ErrorOk = errorOk
addTests(check)
roleVals, issueVals, issueTestStep = oldRoleVals, oldIssueVals, oldIssueTestStep
}
roleVals.AllowAnyName = true
roleVals.EnforceHostnames = true
roleVals.AllowLocalhost = true
roleVals.UseCSRCommonName = true
commonNames.Localhost = true
newOtherNameUtf8 := func(s string) (ret otherNameUtf8) {
pieces := strings.Split(s, ";")
if len(pieces) == 2 {
piecesRest := strings.Split(pieces[1], ":")
if len(piecesRest) == 2 {
switch strings.ToUpper(piecesRest[0]) {
case "UTF-8", "UTF8":
return otherNameUtf8{oid: pieces[0], value: piecesRest[1]}
}
}
}
t.Fatalf("error parsing otherName: %q", s)
return
}
oid1 := "1.3.6.1.4.1.311.20.2.3"
oth1str := oid1 + ";utf8:devops@nope.com"
oth1 := newOtherNameUtf8(oth1str)
oth2 := otherNameUtf8{oid1, "me@example.com"}
// allowNone, allowAll := []string{}, []string{oid1 + ";UTF-8:*"}
allowNone, allowAll := []string{}, []string{"*"}
// OtherSANs not allowed and not provided, should not be an error.
addOtherSANTests(useCSRs, false, allowNone, false, nil, nil, getOtherCheck())
// OtherSANs not allowed, valid OtherSANs provided, should be an error.
addOtherSANTests(useCSRs, false, allowNone, true, []string{oth1str}, nil, nil)
// OtherSANs allowed, bogus OtherSANs provided, should be an error.
addOtherSANTests(useCSRs, false, allowAll, true, []string{"foobar"}, nil, nil)
// Given OtherSANs as API argument and useCSRSANs false, CSR arg ignored.
addOtherSANTests(useCSRs, false, allowAll, false, []string{oth1str},
[]otherNameUtf8{oth2}, getOtherCheck(oth1))
if useCSRs {
// OtherSANs not allowed, valid OtherSANs provided via CSR, should be an error.
addOtherSANTests(useCSRs, true, allowNone, true, nil, []otherNameUtf8{oth1}, nil)
// Given OtherSANs as both API and CSR arguments and useCSRSANs=true, API arg ignored.
addOtherSANTests(useCSRs, false, allowAll, false, []string{oth2.String()},
[]otherNameUtf8{oth1}, getOtherCheck(oth2))
}
}
// Lease tests
{
roleTestStep.ErrorOk = true
roleVals.Lease = ""
roleVals.MaxTTL = 0
addTests(nil)
roleVals.Lease = "12h"
roleVals.MaxTTL = 6 * time.Hour
addTests(nil)
roleTestStep.ErrorOk = false
2018-05-09 16:47:00 +00:00
roleVals.TTL = 0
roleVals.MaxTTL = 12 * time.Hour
}
2016-01-28 20:18:07 +00:00
// Listing test
ret = append(ret, logicaltest.TestStep{
Operation: logical.ListOperation,
Path: "roles/",
Check: func(resp *logical.Response) error {
if resp.Data == nil {
return fmt.Errorf("nil data")
}
keysRaw, ok := resp.Data["keys"]
if !ok {
return fmt.Errorf("no keys found")
}
keys, ok := keysRaw.([]string)
if !ok {
return fmt.Errorf("could not convert keys to a string list")
}
if len(keys) != 1 {
return fmt.Errorf("unexpected keys length of %d", len(keys))
}
if keys[0] != "test" {
2016-04-13 19:38:29 +00:00
return fmt.Errorf("unexpected key value of %s", keys[0])
2016-01-28 20:18:07 +00:00
}
return nil
},
})
return ret
}
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
func TestRolesAltIssuer(t *testing.T) {
t.Parallel()
b, s := CreateBackendWithStorage(t)
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
// Create two issuers.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err := CBWrite(b, s, "root/generate/internal", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "root a - example.com",
"issuer_name": "root-a",
"key_type": "ec",
})
require.NoError(t, err)
require.NotNil(t, resp)
rootAPem := resp.Data["certificate"].(string)
rootACert := parseCert(t, rootAPem)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "root/generate/internal", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "root b - example.com",
"issuer_name": "root-b",
"key_type": "ec",
})
require.NoError(t, err)
require.NotNil(t, resp)
rootBPem := resp.Data["certificate"].(string)
rootBCert := parseCert(t, rootBPem)
// Create three roles: one with no assignment, one with explicit root-a,
// one with explicit root-b.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "roles/use-default", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"allow_any_name": true,
"enforce_hostnames": false,
"key_type": "ec",
})
require.NoError(t, err)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "roles/use-root-a", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"allow_any_name": true,
"enforce_hostnames": false,
"key_type": "ec",
"issuer_ref": "root-a",
})
require.NoError(t, err)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "roles/use-root-b", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"allow_any_name": true,
"enforce_hostnames": false,
"issuer_ref": "root-b",
})
require.NoError(t, err)
// Now issue certs against these roles.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "issue/use-default", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "testing",
"ttl": "5s",
})
require.NoError(t, err)
leafPem := resp.Data["certificate"].(string)
leafCert := parseCert(t, leafPem)
err = leafCert.CheckSignatureFrom(rootACert)
require.NoError(t, err, "should be signed by root-a but wasn't")
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "issue/use-root-a", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "testing",
"ttl": "5s",
})
require.NoError(t, err)
leafPem = resp.Data["certificate"].(string)
leafCert = parseCert(t, leafPem)
err = leafCert.CheckSignatureFrom(rootACert)
require.NoError(t, err, "should be signed by root-a but wasn't")
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "issue/use-root-b", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "testing",
"ttl": "5s",
})
require.NoError(t, err)
leafPem = resp.Data["certificate"].(string)
leafCert = parseCert(t, leafPem)
err = leafCert.CheckSignatureFrom(rootBCert)
require.NoError(t, err, "should be signed by root-b but wasn't")
// Update the default issuer to be root B and make sure that the
// use-default role updates.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "config/issuers", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"default": "root-b",
})
require.NoError(t, err)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "issue/use-default", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "testing",
"ttl": "5s",
})
require.NoError(t, err)
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
leafPem = resp.Data["certificate"].(string)
leafCert = parseCert(t, leafPem)
err = leafCert.CheckSignatureFrom(rootBCert)
require.NoError(t, err, "should be signed by root-b but wasn't")
}
func TestBackend_PathFetchValidRaw(t *testing.T) {
t.Parallel()
b, storage := CreateBackendWithStorage(t)
resp, err := b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "root/generate/internal",
Storage: storage,
Data: map[string]interface{}{
"common_name": "test.com",
"ttl": "6h",
},
MountPoint: "pki/",
})
require.NoError(t, err)
if resp != nil && resp.IsError() {
t.Fatalf("failed to generate root, %#v", resp)
}
rootCaAsPem := resp.Data["certificate"].(string)
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
// Chain should contain the root.
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.ReadOperation,
Path: "ca_chain",
Storage: storage,
Data: map[string]interface{}{},
MountPoint: "pki/",
})
require.NoError(t, err)
if resp != nil && resp.IsError() {
t.Fatalf("failed read ca_chain, %#v", resp)
}
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
if strings.Count(string(resp.Data[logical.HTTPRawBody].([]byte)), rootCaAsPem) != 1 {
t.Fatalf("expected raw chain to contain the root cert")
}
// The ca/pem should return us the actual CA...
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.ReadOperation,
Path: "ca/pem",
Storage: storage,
Data: map[string]interface{}{},
MountPoint: "pki/",
})
require.NoError(t, err)
if resp != nil && resp.IsError() {
t.Fatalf("failed read ca/pem, %#v", resp)
}
// check the raw cert matches the response body
if !bytes.Equal(resp.Data[logical.HTTPRawBody].([]byte), []byte(rootCaAsPem)) {
t.Fatalf("failed to get raw cert")
}
_, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "roles/example",
Storage: storage,
Data: map[string]interface{}{
"allowed_domains": "example.com",
"allow_subdomains": "true",
"max_ttl": "1h",
"no_store": "false",
},
MountPoint: "pki/",
})
require.NoError(t, err, "error setting up pki role: %v", err)
// Now issue a short-lived certificate from our pki-external.
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "issue/example",
Storage: storage,
Data: map[string]interface{}{
"common_name": "test.example.com",
"ttl": "5m",
},
MountPoint: "pki/",
})
require.NoError(t, err, "error issuing certificate: %v", err)
require.NotNil(t, resp, "got nil response from issuing request")
issueCrtAsPem := resp.Data["certificate"].(string)
issuedCrt := parseCert(t, issueCrtAsPem)
expectedSerial := serialFromCert(issuedCrt)
expectedCert := []byte(issueCrtAsPem)
// get der cert
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.ReadOperation,
Path: fmt.Sprintf("cert/%s/raw", expectedSerial),
Storage: storage,
})
if resp != nil && resp.IsError() {
t.Fatalf("failed to get raw cert, %#v", resp)
}
if err != nil {
t.Fatal(err)
}
// check the raw cert matches the response body
rawBody := resp.Data[logical.HTTPRawBody].([]byte)
bodyAsPem := []byte(strings.TrimSpace(string(pem.EncodeToMemory(&pem.Block{Type: "CERTIFICATE", Bytes: rawBody}))))
if !bytes.Equal(bodyAsPem, expectedCert) {
t.Fatalf("failed to get raw cert for serial number: %s", expectedSerial)
}
if resp.Data[logical.HTTPContentType] != "application/pkix-cert" {
t.Fatalf("failed to get raw cert content-type")
}
// get pem
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.ReadOperation,
Path: fmt.Sprintf("cert/%s/raw/pem", expectedSerial),
Storage: storage,
})
if resp != nil && resp.IsError() {
t.Fatalf("failed to get raw, %#v", resp)
}
if err != nil {
t.Fatal(err)
}
// check the pem cert matches the response body
if !bytes.Equal(resp.Data[logical.HTTPRawBody].([]byte), expectedCert) {
t.Fatalf("failed to get pem cert")
}
if resp.Data[logical.HTTPContentType] != "application/pem-certificate-chain" {
t.Fatalf("failed to get raw cert content-type")
}
}
2016-06-08 15:46:58 +00:00
func TestBackend_PathFetchCertList(t *testing.T) {
t.Parallel()
2016-06-08 15:46:58 +00:00
// create the backend
b, storage := CreateBackendWithStorage(t)
2016-06-08 15:46:58 +00:00
// generate root
rootData := map[string]interface{}{
"common_name": "test.com",
"ttl": "6h",
}
resp, err := b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "root/generate/internal",
Storage: storage,
Data: rootData,
MountPoint: "pki/",
2016-06-08 15:46:58 +00:00
})
if resp != nil && resp.IsError() {
t.Fatalf("failed to generate root, %#v", resp)
}
if err != nil {
t.Fatal(err)
}
// config urls
urlsData := map[string]interface{}{
"issuing_certificates": "http://127.0.0.1:8200/v1/pki/ca",
"crl_distribution_points": "http://127.0.0.1:8200/v1/pki/crl",
}
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "config/urls",
Storage: storage,
Data: urlsData,
MountPoint: "pki/",
2016-06-08 15:46:58 +00:00
})
if resp != nil && resp.IsError() {
t.Fatalf("failed to config urls, %#v", resp)
}
if err != nil {
t.Fatal(err)
}
// create a role entry
roleData := map[string]interface{}{
"allowed_domains": "test.com",
"allow_subdomains": "true",
"max_ttl": "4h",
}
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "roles/test-example",
Storage: storage,
Data: roleData,
MountPoint: "pki/",
2016-06-08 15:46:58 +00:00
})
if resp != nil && resp.IsError() {
t.Fatalf("failed to create a role, %#v", resp)
}
if err != nil {
t.Fatal(err)
}
2016-06-08 16:49:10 +00:00
// issue some certs
i := 1
for i < 10 {
certData := map[string]interface{}{
"common_name": "example.test.com",
}
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "issue/test-example",
Storage: storage,
Data: certData,
MountPoint: "pki/",
2016-06-08 16:49:10 +00:00
})
if resp != nil && resp.IsError() {
t.Fatalf("failed to issue a cert, %#v", resp)
}
if err != nil {
t.Fatal(err)
}
i = i + 1
2016-06-08 15:46:58 +00:00
}
2016-06-08 16:49:10 +00:00
// list certs
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.ListOperation,
Path: "certs",
Storage: storage,
MountPoint: "pki/",
2016-06-08 15:46:58 +00:00
})
if resp != nil && resp.IsError() {
2016-06-08 16:49:10 +00:00
t.Fatalf("failed to list certs, %#v", resp)
2016-06-08 15:46:58 +00:00
}
if err != nil {
t.Fatal(err)
}
2016-06-08 16:49:10 +00:00
// check that the root and 9 additional certs are all listed
if len(resp.Data["keys"].([]string)) != 10 {
t.Fatalf("failed to list all 10 certs")
}
2016-06-08 15:46:58 +00:00
2016-06-08 16:49:10 +00:00
// list certs/
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.ListOperation,
Path: "certs/",
Storage: storage,
MountPoint: "pki/",
2016-06-08 15:46:58 +00:00
})
if resp != nil && resp.IsError() {
t.Fatalf("failed to list certs, %#v", resp)
}
if err != nil {
t.Fatal(err)
}
2016-06-08 16:49:10 +00:00
// check that the root and 9 additional certs are all listed
if len(resp.Data["keys"].([]string)) != 10 {
t.Fatalf("failed to list all 10 certs")
}
2016-06-08 15:46:58 +00:00
}
func TestBackend_SignVerbatim(t *testing.T) {
t.Parallel()
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
testCases := []struct {
testName string
keyType string
}{
{testName: "RSA", keyType: "rsa"},
{testName: "ED25519", keyType: "ed25519"},
{testName: "EC", keyType: "ec"},
{testName: "Any", keyType: "any"},
}
for _, tc := range testCases {
tc := tc
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
t.Run(tc.testName, func(t *testing.T) {
runTestSignVerbatim(t, tc.keyType)
})
}
}
func runTestSignVerbatim(t *testing.T, keyType string) {
// create the backend
b, storage := CreateBackendWithStorage(t)
// generate root
rootData := map[string]interface{}{
"common_name": "test.com",
"not_after": "9999-12-31T23:59:59Z",
}
resp, err := b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "root/generate/internal",
Storage: storage,
Data: rootData,
MountPoint: "pki/",
})
if resp != nil && resp.IsError() {
t.Fatalf("failed to generate root, %#v", *resp)
}
if err != nil {
t.Fatal(err)
}
// create a CSR and key
key, err := rsa.GenerateKey(rand.Reader, 2048)
if err != nil {
t.Fatal(err)
}
csrReq := &x509.CertificateRequest{
Subject: pkix.Name{
CommonName: "foo.bar.com",
},
// Check that otherName extensions are not duplicated (see hashicorp/vault#16700).
// If these extensions are duplicated, sign-verbatim will fail when parsing the signed certificate on Go 1.19+ (see golang/go#50988).
// On older versions of Go this test will fail due to an explicit check for duplicate otherNames later in this test.
ExtraExtensions: []pkix.Extension{
{
Id: oidExtensionSubjectAltName,
Critical: false,
Value: []byte{0x30, 0x26, 0xA0, 0x24, 0x06, 0x0A, 0x2B, 0x06, 0x01, 0x04, 0x01, 0x82, 0x37, 0x14, 0x02, 0x03, 0xA0, 0x16, 0x0C, 0x14, 0x75, 0x73, 0x65, 0x72, 0x6E, 0x61, 0x6D, 0x65, 0x40, 0x65, 0x78, 0x61, 0x6D, 0x70, 0x6C, 0x65, 0x2E, 0x63, 0x6F, 0x6D},
},
},
}
csr, err := x509.CreateCertificateRequest(rand.Reader, csrReq, key)
if err != nil {
t.Fatal(err)
}
if len(csr) == 0 {
t.Fatal("generated csr is empty")
}
pemCSR := strings.TrimSpace(string(pem.EncodeToMemory(&pem.Block{
Type: "CERTIFICATE REQUEST",
Bytes: csr,
})))
if len(pemCSR) == 0 {
t.Fatal("pem csr is empty")
}
signVerbatimData := map[string]interface{}{
"csr": pemCSR,
}
if keyType == "rsa" {
signVerbatimData["signature_bits"] = 512
}
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "sign-verbatim",
Storage: storage,
Data: signVerbatimData,
MountPoint: "pki/",
})
if resp != nil && resp.IsError() {
t.Fatalf("failed to sign-verbatim basic CSR: %#v", *resp)
}
if err != nil {
t.Fatal(err)
}
if resp.Secret != nil {
t.Fatal("secret is not nil")
}
// create a role entry; we use this to check that sign-verbatim when used with a role is still honoring TTLs
roleData := map[string]interface{}{
"ttl": "4h",
"max_ttl": "8h",
"key_type": keyType,
"not_before_duration": "2h",
}
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "roles/test",
Storage: storage,
Data: roleData,
MountPoint: "pki/",
})
if resp != nil && resp.IsError() {
t.Fatalf("failed to create a role, %#v", *resp)
}
if err != nil {
t.Fatal(err)
}
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "sign-verbatim/test",
Storage: storage,
Data: map[string]interface{}{
"csr": pemCSR,
"ttl": "5h",
},
MountPoint: "pki/",
})
if resp != nil && resp.IsError() {
t.Fatalf("failed to sign-verbatim ttl'd CSR: %#v", *resp)
}
if err != nil {
t.Fatal(err)
}
if resp.Secret != nil {
t.Fatal("got a lease when we should not have")
}
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "sign-verbatim/test",
Storage: storage,
Data: map[string]interface{}{
"csr": pemCSR,
"ttl": "12h",
},
MountPoint: "pki/",
})
if err != nil {
t.Fatal(err)
}
if resp != nil && resp.IsError() {
t.Fatalf(resp.Error().Error())
}
if resp.Data == nil || resp.Data["certificate"] == nil {
t.Fatal("did not get expected data")
}
certString := resp.Data["certificate"].(string)
block, _ := pem.Decode([]byte(certString))
if block == nil {
t.Fatal("nil pem block")
}
certs, err := x509.ParseCertificates(block.Bytes)
if err != nil {
t.Fatal(err)
}
if len(certs) != 1 {
t.Fatalf("expected a single cert, got %d", len(certs))
}
cert := certs[0]
if math.Abs(float64(time.Now().Add(12*time.Hour).Unix()-cert.NotAfter.Unix())) < 10 {
t.Fatalf("sign-verbatim did not properly cap validity period (notAfter) on signed CSR: was %v vs requested %v but should've been %v", cert.NotAfter, time.Now().Add(12*time.Hour), time.Now().Add(8*time.Hour))
}
if math.Abs(float64(time.Now().Add(-2*time.Hour).Unix()-cert.NotBefore.Unix())) > 10 {
t.Fatalf("sign-verbatim did not properly cap validity period (notBefore) on signed CSR: was %v vs expected %v", cert.NotBefore, time.Now().Add(-2*time.Hour))
}
// Now check signing a certificate using the not_after input using the Y10K value
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "sign-verbatim/test",
Storage: storage,
Data: map[string]interface{}{
"csr": pemCSR,
"not_after": "9999-12-31T23:59:59Z",
},
MountPoint: "pki/",
})
if err != nil {
t.Fatal(err)
}
if resp != nil && resp.IsError() {
t.Fatalf(resp.Error().Error())
}
if resp.Data == nil || resp.Data["certificate"] == nil {
t.Fatal("did not get expected data")
}
certString = resp.Data["certificate"].(string)
block, _ = pem.Decode([]byte(certString))
if block == nil {
t.Fatal("nil pem block")
}
certs, err = x509.ParseCertificates(block.Bytes)
if err != nil {
t.Fatal(err)
}
if len(certs) != 1 {
t.Fatalf("expected a single cert, got %d", len(certs))
}
cert = certs[0]
// Fallback check for duplicate otherName, necessary on Go versions before 1.19.
// We assume that there is only one SAN in the original CSR and that it is an otherName.
san_count := 0
for _, ext := range cert.Extensions {
if ext.Id.Equal(oidExtensionSubjectAltName) {
san_count += 1
}
}
if san_count != 1 {
t.Fatalf("expected one SAN extension, got %d", san_count)
}
notAfter := cert.NotAfter.Format(time.RFC3339)
if notAfter != "9999-12-31T23:59:59Z" {
t.Fatal(fmt.Errorf("not after from certificate is not matching with input parameter"))
}
// now check that if we set generate-lease it takes it from the role and the TTLs match
roleData = map[string]interface{}{
"ttl": "4h",
"max_ttl": "8h",
"generate_lease": true,
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
"key_type": keyType,
}
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "roles/test",
Storage: storage,
Data: roleData,
MountPoint: "pki/",
})
if resp != nil && resp.IsError() {
t.Fatalf("failed to create a role, %#v", *resp)
}
if err != nil {
t.Fatal(err)
}
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "sign-verbatim/test",
Storage: storage,
Data: map[string]interface{}{
"csr": pemCSR,
"ttl": "5h",
},
MountPoint: "pki/",
})
if resp != nil && resp.IsError() {
t.Fatalf("failed to sign-verbatim role-leased CSR: %#v", *resp)
}
if err != nil {
t.Fatal(err)
}
if resp.Secret == nil {
t.Fatalf("secret is nil, response is %#v", *resp)
}
if math.Abs(float64(resp.Secret.TTL-(5*time.Hour))) > float64(5*time.Hour) {
t.Fatalf("ttl not default; wanted %v, got %v", b.System().DefaultLeaseTTL(), resp.Secret.TTL)
}
}
2018-03-20 18:54:10 +00:00
func TestBackend_Root_Idempotency(t *testing.T) {
t.Parallel()
b, s := CreateBackendWithStorage(t)
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
// This is a change within 1.11, we are no longer idempotent across generate/internal calls.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err := CBWrite(b, s, "root/generate/internal", map[string]interface{}{
"common_name": "myvault.com",
})
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
require.NoError(t, err)
require.NotNil(t, resp, "expected ca info")
keyId1 := resp.Data["key_id"]
issuerId1 := resp.Data["issuer_id"]
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBRead(b, s, "cert/ca_chain")
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
require.NoError(t, err, "error reading ca_chain: %v", err)
r1Data := resp.Data
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
// Calling generate/internal should generate a new CA as well.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "root/generate/internal", map[string]interface{}{
"common_name": "myvault.com",
})
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
require.NoError(t, err)
require.NotNil(t, resp, "expected ca info")
keyId2 := resp.Data["key_id"]
issuerId2 := resp.Data["issuer_id"]
// Make sure that we actually generated different issuer and key values
require.NotEqual(t, keyId1, keyId2)
require.NotEqual(t, issuerId1, issuerId2)
// Now because the issued CA's have no links, the call to ca_chain should return the same data (ca chain from default)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBRead(b, s, "cert/ca_chain")
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
require.NoError(t, err, "error reading ca_chain: %v", err)
r2Data := resp.Data
if !reflect.DeepEqual(r1Data, r2Data) {
t.Fatal("got different ca certs")
}
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
// Now let's validate that the import bundle is idempotent.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
pemBundleRootCA := rootCACertPEM + "\n" + rootCAKeyPEM
resp, err = CBWrite(b, s, "config/ca", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"pem_bundle": pemBundleRootCA,
})
require.NoError(t, err)
require.NotNil(t, resp, "expected ca info")
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
firstImportedKeys := resp.Data["imported_keys"].([]string)
firstImportedIssuers := resp.Data["imported_issuers"].([]string)
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
require.NotContains(t, firstImportedKeys, keyId1)
require.NotContains(t, firstImportedKeys, keyId2)
require.NotContains(t, firstImportedIssuers, issuerId1)
require.NotContains(t, firstImportedIssuers, issuerId2)
// Performing this again should result in no key/issuer ids being imported/generated.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "config/ca", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"pem_bundle": pemBundleRootCA,
})
require.NoError(t, err)
require.NotNil(t, resp, "expected ca info")
secondImportedKeys := resp.Data["imported_keys"]
secondImportedIssuers := resp.Data["imported_issuers"]
require.Nil(t, secondImportedKeys)
require.Nil(t, secondImportedIssuers)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBDelete(b, s, "root")
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
require.NoError(t, err)
require.NotNil(t, resp)
require.Equal(t, 1, len(resp.Warnings))
// Make sure we can delete twice...
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBDelete(b, s, "root")
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
require.NoError(t, err)
require.NotNil(t, resp)
require.Equal(t, 1, len(resp.Warnings))
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBRead(b, s, "cert/ca_chain")
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
require.Error(t, err, "expected an error fetching deleted ca_chain")
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
// We should be able to import the same ca bundle as before and get a different key/issuer ids
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "config/ca", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"pem_bundle": pemBundleRootCA,
})
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
require.NoError(t, err)
require.NotNil(t, resp, "expected ca info")
postDeleteImportedKeys := resp.Data["imported_keys"]
postDeleteImportedIssuers := resp.Data["imported_issuers"]
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
// Make sure that we actually generated different issuer and key values, then the previous import
require.NotNil(t, postDeleteImportedKeys)
require.NotNil(t, postDeleteImportedIssuers)
require.NotEqual(t, postDeleteImportedKeys, firstImportedKeys)
require.NotEqual(t, postDeleteImportedIssuers, firstImportedIssuers)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBRead(b, s, "cert/ca_chain")
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
require.NoError(t, err)
caChainPostDelete := resp.Data
if reflect.DeepEqual(r1Data, caChainPostDelete) {
t.Fatal("ca certs from ca_chain were the same post delete, should have changed.")
}
}
func TestBackend_SignIntermediate_AllowedPastCA(t *testing.T) {
t.Parallel()
b_root, s_root := CreateBackendWithStorage(t)
b_int, s_int := CreateBackendWithStorage(t)
var err error
// Direct issuing from root
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b_root, s_root, "root/generate/internal", map[string]interface{}{
"ttl": "40h",
"common_name": "myvault.com",
})
if err != nil {
t.Fatal(err)
}
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b_root, s_root, "roles/test", map[string]interface{}{
"allow_bare_domains": true,
"allow_subdomains": true,
})
if err != nil {
t.Fatal(err)
}
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err := CBWrite(b_int, s_int, "intermediate/generate/internal", map[string]interface{}{
"common_name": "myint.com",
})
if err != nil {
t.Fatal(err)
}
csr := resp.Data["csr"]
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b_root, s_root, "sign/test", map[string]interface{}{
"common_name": "myint.com",
"csr": csr,
"ttl": "60h",
})
if err == nil {
t.Fatal("expected error")
}
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b_root, s_root, "sign-verbatim/test", map[string]interface{}{
"common_name": "myint.com",
"other_sans": "1.3.6.1.4.1.311.20.2.3;utf8:caadmin@example.com",
"csr": csr,
"ttl": "60h",
})
if err == nil {
t.Fatal("expected error")
}
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b_root, s_root, "root/sign-intermediate", map[string]interface{}{
"common_name": "myint.com",
"other_sans": "1.3.6.1.4.1.311.20.2.3;utf8:caadmin@example.com",
"csr": csr,
"ttl": "60h",
})
if err != nil {
t.Fatalf("got error: %v", err)
}
if resp == nil {
t.Fatal("got nil response")
}
if len(resp.Warnings) == 0 {
t.Fatalf("expected warnings, got %#v", *resp)
}
}
func TestBackend_ConsulSignLeafWithLegacyRole(t *testing.T) {
t.Parallel()
// create the backend
b, s := CreateBackendWithStorage(t)
// generate root
data, err := CBWrite(b, s, "root/generate/internal", map[string]interface{}{
"ttl": "40h",
"common_name": "myvault.com",
})
require.NoError(t, err, "failed generating internal root cert")
rootCaPem := data.Data["certificate"].(string)
// Create a signing role like Consul did with the default args prior to Vault 1.10
_, err = CBWrite(b, s, "roles/test", map[string]interface{}{
"allow_any_name": true,
"allowed_serial_numbers": []string{"MySerialNumber"},
"key_type": "any",
"key_bits": "2048",
"signature_bits": "256",
})
require.NoError(t, err, "failed creating legacy role")
_, csrPem := generateTestCsr(t, certutil.ECPrivateKey, 256)
data, err = CBWrite(b, s, "sign/test", map[string]interface{}{
"csr": csrPem,
})
require.NoError(t, err, "failed signing csr")
certAsPem := data.Data["certificate"].(string)
signedCert := parseCert(t, certAsPem)
rootCert := parseCert(t, rootCaPem)
requireSignedBy(t, signedCert, rootCert)
}
func TestBackend_SignSelfIssued(t *testing.T) {
t.Parallel()
// create the backend
b, storage := CreateBackendWithStorage(t)
// generate root
rootData := map[string]interface{}{
"common_name": "test.com",
"ttl": "172800",
}
resp, err := b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "root/generate/internal",
Storage: storage,
Data: rootData,
MountPoint: "pki/",
})
if resp != nil && resp.IsError() {
t.Fatalf("failed to generate root, %#v", *resp)
}
if err != nil {
t.Fatal(err)
}
key, err := rsa.GenerateKey(rand.Reader, 2048)
if err != nil {
t.Fatal(err)
}
template := &x509.Certificate{
Subject: pkix.Name{
CommonName: "foo.bar.com",
},
SerialNumber: big.NewInt(1234),
IsCA: false,
BasicConstraintsValid: true,
}
ss, _ := getSelfSigned(t, template, template, key)
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "root/sign-self-issued",
Storage: storage,
Data: map[string]interface{}{
"certificate": ss,
},
MountPoint: "pki/",
})
if err != nil {
t.Fatal(err)
}
if resp == nil {
t.Fatal("got nil response")
}
if !resp.IsError() {
t.Fatalf("expected error due to non-CA; got: %#v", *resp)
}
// Set CA to true, but leave issuer alone
template.IsCA = true
issuer := &x509.Certificate{
Subject: pkix.Name{
CommonName: "bar.foo.com",
},
SerialNumber: big.NewInt(2345),
IsCA: true,
BasicConstraintsValid: true,
}
ss, ssCert := getSelfSigned(t, template, issuer, key)
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "root/sign-self-issued",
Storage: storage,
Data: map[string]interface{}{
"certificate": ss,
},
MountPoint: "pki/",
})
if err != nil {
t.Fatal(err)
}
if resp == nil {
t.Fatal("got nil response")
}
if !resp.IsError() {
t.Fatalf("expected error due to different issuer; cert info is\nIssuer\n%#v\nSubject\n%#v\n", ssCert.Issuer, ssCert.Subject)
}
ss, _ = getSelfSigned(t, template, template, key)
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "root/sign-self-issued",
Storage: storage,
Data: map[string]interface{}{
"certificate": ss,
},
MountPoint: "pki/",
})
if err != nil {
t.Fatal(err)
}
if resp == nil {
t.Fatal("got nil response")
}
if resp.IsError() {
t.Fatalf("error in response: %s", resp.Error().Error())
}
newCertString := resp.Data["certificate"].(string)
block, _ := pem.Decode([]byte(newCertString))
newCert, err := x509.ParseCertificate(block.Bytes)
if err != nil {
t.Fatal(err)
}
sc := b.makeStorageContext(context.Background(), storage)
signingBundle, err := sc.fetchCAInfo(defaultRef, ReadOnlyUsage)
if err != nil {
t.Fatal(err)
}
if reflect.DeepEqual(newCert.Subject, newCert.Issuer) {
t.Fatal("expected different subject/issuer")
}
if !reflect.DeepEqual(newCert.Issuer, signingBundle.Certificate.Subject) {
t.Fatalf("expected matching issuer/CA subject\n\nIssuer:\n%#v\nSubject:\n%#v\n", newCert.Issuer, signingBundle.Certificate.Subject)
}
if bytes.Equal(newCert.AuthorityKeyId, newCert.SubjectKeyId) {
t.Fatal("expected different authority/subject")
}
if !bytes.Equal(newCert.AuthorityKeyId, signingBundle.Certificate.SubjectKeyId) {
t.Fatal("expected authority on new cert to be same as signing subject")
}
if newCert.Subject.CommonName != "foo.bar.com" {
t.Fatalf("unexpected common name on new cert: %s", newCert.Subject.CommonName)
}
}
// TestBackend_SignSelfIssued_DifferentTypes tests the functionality of the
// require_matching_certificate_algorithms flag.
func TestBackend_SignSelfIssued_DifferentTypes(t *testing.T) {
t.Parallel()
// create the backend
b, storage := CreateBackendWithStorage(t)
// generate root
rootData := map[string]interface{}{
"common_name": "test.com",
"ttl": "172800",
"key_type": "ec",
"key_bits": "521",
}
resp, err := b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "root/generate/internal",
Storage: storage,
Data: rootData,
MountPoint: "pki/",
})
if resp != nil && resp.IsError() {
t.Fatalf("failed to generate root, %#v", *resp)
}
if err != nil {
t.Fatal(err)
}
key, err := rsa.GenerateKey(rand.Reader, 2048)
if err != nil {
t.Fatal(err)
}
template := &x509.Certificate{
Subject: pkix.Name{
CommonName: "foo.bar.com",
},
2018-11-07 21:52:01 +00:00
SerialNumber: big.NewInt(1234),
IsCA: true,
BasicConstraintsValid: true,
}
// Tests absent the flag
ss, _ := getSelfSigned(t, template, template, key)
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "root/sign-self-issued",
Storage: storage,
Data: map[string]interface{}{
"certificate": ss,
},
MountPoint: "pki/",
})
if err != nil {
t.Fatal(err)
}
if resp == nil {
t.Fatal("got nil response")
}
// Set CA to true, but leave issuer alone
template.IsCA = true
// Tests with flag present but false
ss, _ = getSelfSigned(t, template, template, key)
resp, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "root/sign-self-issued",
Storage: storage,
Data: map[string]interface{}{
"certificate": ss,
"require_matching_certificate_algorithms": false,
},
MountPoint: "pki/",
})
if err != nil {
t.Fatal(err)
}
if resp == nil {
t.Fatal("got nil response")
}
// Test with flag present and true
ss, _ = getSelfSigned(t, template, template, key)
_, err = b.HandleRequest(context.Background(), &logical.Request{
Operation: logical.UpdateOperation,
Path: "root/sign-self-issued",
Storage: storage,
Data: map[string]interface{}{
"certificate": ss,
"require_matching_certificate_algorithms": true,
},
MountPoint: "pki/",
})
if err == nil {
t.Fatal("expected error due to mismatched algorithms")
}
}
2018-02-16 22:19:34 +00:00
// This is a really tricky test because the Go stdlib asn1 package is incapable
// of doing the right thing with custom OID SANs (see comments in the package,
// it's readily admitted that it's too magic) but that means that any
// validation logic written for this test isn't being independently verified,
// as in, if cryptobytes is used to decode it to make the test work, that
// doesn't mean we're encoding and decoding correctly, only that we made the
// test pass. Instead, when run verbosely it will first perform a bunch of
// checks to verify that the OID SAN logic doesn't screw up other SANs, then
// will spit out the PEM. This can be validated independently.
//
// You want the hex dump of the octet string corresponding to the X509v3
// Subject Alternative Name. There's a nice online utility at
// https://lapo.it/asn1js that can be used to view the structure of an
// openssl-generated other SAN at
// https://lapo.it/asn1js/#3022A020060A2B060104018237140203A0120C106465766F7073406C6F63616C686F7374
// (openssl asn1parse can also be used with -strparse using an offset of the
// hex blob for the subject alternative names extension).
//
// The structure output from here should match that precisely (even if the OID
// itself doesn't) in the second test.
//
// The test that encodes two should have them be in separate elements in the
// top-level sequence; see
// https://lapo.it/asn1js/#3046A020060A2B060104018237140203A0120C106465766F7073406C6F63616C686F7374A022060A2B060104018237140204A0140C12322D6465766F7073406C6F63616C686F7374 for an openssl-generated example.
//
// The good news is that it's valid to simply copy and paste the PEM output from
2018-02-16 22:19:34 +00:00
// here into the form at that site as it will do the right thing so it's pretty
// easy to validate.
func TestBackend_OID_SANs(t *testing.T) {
t.Parallel()
b, s := CreateBackendWithStorage(t)
2018-02-16 22:19:34 +00:00
var err error
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
var resp *logical.Response
2018-02-16 22:19:34 +00:00
var certStr string
var block *pem.Block
var cert *x509.Certificate
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "root/generate/internal", map[string]interface{}{
2018-02-16 22:19:34 +00:00
"ttl": "40h",
"common_name": "myvault.com",
})
if err != nil {
t.Fatal(err)
}
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "roles/test", map[string]interface{}{
2018-02-16 22:19:34 +00:00
"allowed_domains": []string{"foobar.com", "zipzap.com"},
"allow_bare_domains": true,
"allow_subdomains": true,
"allow_ip_sans": true,
"allowed_other_sans": "1.3.6.1.4.1.311.20.2.3;UTF8:devops@*,1.3.6.1.4.1.311.20.2.4;utf8:d*e@foobar.com",
})
if err != nil {
t.Fatal(err)
}
// Get a baseline before adding OID SANs. In the next sections we'll verify
// that the SANs are all added even as the OID SAN inclusion forces other
// adding logic (custom rather than built-in Golang logic)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "issue/test", map[string]interface{}{
2018-02-16 22:19:34 +00:00
"common_name": "foobar.com",
"ip_sans": "1.2.3.4",
"alt_names": "foobar.com,foo.foobar.com,bar.foobar.com",
2018-02-16 22:19:34 +00:00
"ttl": "1h",
})
if err != nil {
t.Fatal(err)
}
certStr = resp.Data["certificate"].(string)
block, _ = pem.Decode([]byte(certStr))
cert, err = x509.ParseCertificate(block.Bytes)
if err != nil {
t.Fatal(err)
}
if cert.IPAddresses[0].String() != "1.2.3.4" {
t.Fatalf("unexpected IP SAN %q", cert.IPAddresses[0].String())
}
if len(cert.DNSNames) != 3 ||
cert.DNSNames[0] != "bar.foobar.com" ||
cert.DNSNames[1] != "foo.foobar.com" ||
cert.DNSNames[2] != "foobar.com" {
2018-02-16 22:19:34 +00:00
t.Fatalf("unexpected DNS SANs %v", cert.DNSNames)
}
// First test some bad stuff that shouldn't work
_, err = CBWrite(b, s, "issue/test", map[string]interface{}{
2018-02-16 22:19:34 +00:00
"common_name": "foobar.com",
"ip_sans": "1.2.3.4",
"alt_names": "foo.foobar.com,bar.foobar.com",
"ttl": "1h",
// Not a valid value for the first possibility
"other_sans": "1.3.6.1.4.1.311.20.2.3;UTF8:devop@nope.com",
})
if err == nil {
t.Fatal("expected error")
}
_, err = CBWrite(b, s, "issue/test", map[string]interface{}{
2018-02-16 22:19:34 +00:00
"common_name": "foobar.com",
"ip_sans": "1.2.3.4",
"alt_names": "foo.foobar.com,bar.foobar.com",
"ttl": "1h",
// Not a valid OID for the first possibility
"other_sans": "1.3.6.1.4.1.311.20.2.5;UTF8:devops@nope.com",
})
if err == nil {
t.Fatal("expected error")
}
_, err = CBWrite(b, s, "issue/test", map[string]interface{}{
2018-02-16 22:19:34 +00:00
"common_name": "foobar.com",
"ip_sans": "1.2.3.4",
"alt_names": "foo.foobar.com,bar.foobar.com",
"ttl": "1h",
// Not a valid name for the second possibility
"other_sans": "1.3.6.1.4.1.311.20.2.4;UTF8:d34g@foobar.com",
})
if err == nil {
t.Fatal("expected error")
}
_, err = CBWrite(b, s, "issue/test", map[string]interface{}{
2018-02-16 22:19:34 +00:00
"common_name": "foobar.com",
"ip_sans": "1.2.3.4",
"alt_names": "foo.foobar.com,bar.foobar.com",
"ttl": "1h",
// Not a valid OID for the second possibility
"other_sans": "1.3.6.1.4.1.311.20.2.5;UTF8:d34e@foobar.com",
})
if err == nil {
t.Fatal("expected error")
}
_, err = CBWrite(b, s, "issue/test", map[string]interface{}{
2018-02-16 22:19:34 +00:00
"common_name": "foobar.com",
"ip_sans": "1.2.3.4",
"alt_names": "foo.foobar.com,bar.foobar.com",
"ttl": "1h",
// Not a valid type
"other_sans": "1.3.6.1.4.1.311.20.2.5;UTF2:d34e@foobar.com",
})
if err == nil {
t.Fatal("expected error")
}
// Valid for first possibility
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "issue/test", map[string]interface{}{
2018-02-16 22:19:34 +00:00
"common_name": "foobar.com",
"ip_sans": "1.2.3.4",
"alt_names": "foo.foobar.com,bar.foobar.com",
"ttl": "1h",
"other_sans": "1.3.6.1.4.1.311.20.2.3;utf8:devops@nope.com",
})
if err != nil {
t.Fatal(err)
}
certStr = resp.Data["certificate"].(string)
block, _ = pem.Decode([]byte(certStr))
cert, err = x509.ParseCertificate(block.Bytes)
if err != nil {
t.Fatal(err)
}
if cert.IPAddresses[0].String() != "1.2.3.4" {
t.Fatalf("unexpected IP SAN %q", cert.IPAddresses[0].String())
}
if len(cert.DNSNames) != 3 ||
cert.DNSNames[0] != "bar.foobar.com" ||
cert.DNSNames[1] != "foo.foobar.com" ||
cert.DNSNames[2] != "foobar.com" {
2018-02-16 22:19:34 +00:00
t.Fatalf("unexpected DNS SANs %v", cert.DNSNames)
}
if len(os.Getenv("VAULT_VERBOSE_PKITESTS")) > 0 {
t.Logf("certificate 1 to check:\n%s", certStr)
}
2018-02-16 22:19:34 +00:00
// Valid for second possibility
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "issue/test", map[string]interface{}{
2018-02-16 22:19:34 +00:00
"common_name": "foobar.com",
"ip_sans": "1.2.3.4",
"alt_names": "foo.foobar.com,bar.foobar.com",
"ttl": "1h",
"other_sans": "1.3.6.1.4.1.311.20.2.4;UTF8:d234e@foobar.com",
})
if err != nil {
t.Fatal(err)
}
certStr = resp.Data["certificate"].(string)
block, _ = pem.Decode([]byte(certStr))
cert, err = x509.ParseCertificate(block.Bytes)
if err != nil {
t.Fatal(err)
}
if cert.IPAddresses[0].String() != "1.2.3.4" {
t.Fatalf("unexpected IP SAN %q", cert.IPAddresses[0].String())
}
if len(cert.DNSNames) != 3 ||
cert.DNSNames[0] != "bar.foobar.com" ||
cert.DNSNames[1] != "foo.foobar.com" ||
cert.DNSNames[2] != "foobar.com" {
2018-02-16 22:19:34 +00:00
t.Fatalf("unexpected DNS SANs %v", cert.DNSNames)
}
if len(os.Getenv("VAULT_VERBOSE_PKITESTS")) > 0 {
t.Logf("certificate 2 to check:\n%s", certStr)
}
2018-02-16 22:19:34 +00:00
// Valid for both
oid1, type1, val1 := "1.3.6.1.4.1.311.20.2.3", "utf8", "devops@nope.com"
oid2, type2, val2 := "1.3.6.1.4.1.311.20.2.4", "utf-8", "d234e@foobar.com"
otherNames := []string{
fmt.Sprintf("%s;%s:%s", oid1, type1, val1),
fmt.Sprintf("%s;%s:%s", oid2, type2, val2),
}
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "issue/test", map[string]interface{}{
2018-02-16 22:19:34 +00:00
"common_name": "foobar.com",
"ip_sans": "1.2.3.4",
"alt_names": "foo.foobar.com,bar.foobar.com",
"ttl": "1h",
"other_sans": strings.Join(otherNames, ","),
2018-02-16 22:19:34 +00:00
})
if err != nil {
t.Fatal(err)
}
certStr = resp.Data["certificate"].(string)
block, _ = pem.Decode([]byte(certStr))
cert, err = x509.ParseCertificate(block.Bytes)
if err != nil {
t.Fatal(err)
}
if cert.IPAddresses[0].String() != "1.2.3.4" {
t.Fatalf("unexpected IP SAN %q", cert.IPAddresses[0].String())
}
if len(cert.DNSNames) != 3 ||
cert.DNSNames[0] != "bar.foobar.com" ||
cert.DNSNames[1] != "foo.foobar.com" ||
cert.DNSNames[2] != "foobar.com" {
2018-02-16 22:19:34 +00:00
t.Fatalf("unexpected DNS SANs %v", cert.DNSNames)
}
expectedOtherNames := []otherNameUtf8{{oid1, val1}, {oid2, val2}}
foundOtherNames, err := getOtherSANsFromX509Extensions(cert.Extensions)
if err != nil {
t.Fatal(err)
}
if diff := deep.Equal(expectedOtherNames, foundOtherNames); len(diff) != 0 {
t.Errorf("unexpected otherNames: %v", diff)
}
if len(os.Getenv("VAULT_VERBOSE_PKITESTS")) > 0 {
t.Logf("certificate 3 to check:\n%s", certStr)
}
}
func TestBackend_AllowedSerialNumbers(t *testing.T) {
t.Parallel()
b, s := CreateBackendWithStorage(t)
var err error
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
var resp *logical.Response
var certStr string
var block *pem.Block
var cert *x509.Certificate
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "root/generate/internal", map[string]interface{}{
"ttl": "40h",
"common_name": "myvault.com",
})
if err != nil {
t.Fatal(err)
}
// First test that Serial Numbers are not allowed
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "roles/test", map[string]interface{}{
"allow_any_name": true,
"enforce_hostnames": false,
})
if err != nil {
t.Fatal(err)
}
_, err = CBWrite(b, s, "issue/test", map[string]interface{}{
"common_name": "foobar",
"ttl": "1h",
})
if err != nil {
t.Fatal(err)
}
_, err = CBWrite(b, s, "issue/test", map[string]interface{}{
"common_name": "foobar",
"ttl": "1h",
"serial_number": "foobar",
})
if err == nil {
t.Fatal("expected error")
}
// Update the role to allow serial numbers
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "roles/test", map[string]interface{}{
"allow_any_name": true,
"enforce_hostnames": false,
"allowed_serial_numbers": "f00*,b4r*",
})
if err != nil {
t.Fatal(err)
}
_, err = CBWrite(b, s, "issue/test", map[string]interface{}{
"common_name": "foobar",
"ttl": "1h",
// Not a valid serial number
"serial_number": "foobar",
})
if err == nil {
t.Fatal("expected error")
}
// Valid for first possibility
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "issue/test", map[string]interface{}{
"common_name": "foobar",
"serial_number": "f00bar",
})
if err != nil {
t.Fatal(err)
}
certStr = resp.Data["certificate"].(string)
block, _ = pem.Decode([]byte(certStr))
cert, err = x509.ParseCertificate(block.Bytes)
if err != nil {
t.Fatal(err)
}
if cert.Subject.SerialNumber != "f00bar" {
t.Fatalf("unexpected Subject SerialNumber %s", cert.Subject.SerialNumber)
}
if len(os.Getenv("VAULT_VERBOSE_PKITESTS")) > 0 {
t.Logf("certificate 1 to check:\n%s", certStr)
}
// Valid for second possibility
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "issue/test", map[string]interface{}{
"common_name": "foobar",
"serial_number": "b4rf00",
})
if err != nil {
t.Fatal(err)
}
certStr = resp.Data["certificate"].(string)
block, _ = pem.Decode([]byte(certStr))
cert, err = x509.ParseCertificate(block.Bytes)
if err != nil {
t.Fatal(err)
}
if cert.Subject.SerialNumber != "b4rf00" {
t.Fatalf("unexpected Subject SerialNumber %s", cert.Subject.SerialNumber)
}
if len(os.Getenv("VAULT_VERBOSE_PKITESTS")) > 0 {
t.Logf("certificate 2 to check:\n%s", certStr)
}
2018-02-16 22:19:34 +00:00
}
2018-06-15 19:32:25 +00:00
func TestBackend_URI_SANs(t *testing.T) {
t.Parallel()
b, s := CreateBackendWithStorage(t)
2018-06-15 19:32:25 +00:00
var err error
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "root/generate/internal", map[string]interface{}{
2018-06-15 19:32:25 +00:00
"ttl": "40h",
"common_name": "myvault.com",
})
if err != nil {
t.Fatal(err)
}
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "roles/test", map[string]interface{}{
2018-06-15 19:32:25 +00:00
"allowed_domains": []string{"foobar.com", "zipzap.com"},
"allow_bare_domains": true,
"allow_subdomains": true,
"allow_ip_sans": true,
"allowed_uri_sans": []string{"http://someuri/abc", "spiffe://host.com/*"},
})
if err != nil {
t.Fatal(err)
}
// First test some bad stuff that shouldn't work
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "issue/test", map[string]interface{}{
2018-06-15 19:32:25 +00:00
"common_name": "foobar.com",
"ip_sans": "1.2.3.4",
"alt_names": "foo.foobar.com,bar.foobar.com",
"ttl": "1h",
"uri_sans": "http://www.mydomain.com/zxf",
})
if err == nil {
t.Fatal("expected error")
}
// Test valid single entry
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "issue/test", map[string]interface{}{
2018-06-15 19:32:25 +00:00
"common_name": "foobar.com",
"ip_sans": "1.2.3.4",
"alt_names": "foo.foobar.com,bar.foobar.com",
"ttl": "1h",
"uri_sans": "http://someuri/abc",
})
if err != nil {
t.Fatal(err)
}
// Test globed entry
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "issue/test", map[string]interface{}{
2018-06-15 19:32:25 +00:00
"common_name": "foobar.com",
"ip_sans": "1.2.3.4",
"alt_names": "foo.foobar.com,bar.foobar.com",
"ttl": "1h",
"uri_sans": "spiffe://host.com/something",
})
if err != nil {
t.Fatal(err)
}
// Test multiple entries
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err := CBWrite(b, s, "issue/test", map[string]interface{}{
2018-06-15 19:32:25 +00:00
"common_name": "foobar.com",
"ip_sans": "1.2.3.4",
"alt_names": "foo.foobar.com,bar.foobar.com",
"ttl": "1h",
"uri_sans": "spiffe://host.com/something,http://someuri/abc",
})
if err != nil {
t.Fatal(err)
}
certStr := resp.Data["certificate"].(string)
block, _ := pem.Decode([]byte(certStr))
cert, err := x509.ParseCertificate(block.Bytes)
if err != nil {
t.Fatal(err)
}
URI0, _ := url.Parse("spiffe://host.com/something")
URI1, _ := url.Parse("http://someuri/abc")
if len(cert.URIs) != 2 {
t.Fatalf("expected 2 valid URIs SANs %v", cert.URIs)
}
if cert.URIs[0].String() != URI0.String() || cert.URIs[1].String() != URI1.String() {
t.Fatalf(
"expected URIs SANs %v to equal provided values spiffe://host.com/something, http://someuri/abc",
cert.URIs)
}
}
func TestBackend_AllowedURISANsTemplate(t *testing.T) {
t.Parallel()
coreConfig := &vault.CoreConfig{
CredentialBackends: map[string]logical.Factory{
"userpass": userpass.Factory,
},
LogicalBackends: map[string]logical.Factory{
"pki": Factory,
},
}
cluster := vault.NewTestCluster(t, coreConfig, &vault.TestClusterOptions{
HandlerFunc: vaulthttp.Handler,
})
cluster.Start()
defer cluster.Cleanup()
client := cluster.Cores[0].Client
// Write test policy for userpass auth method.
err := client.Sys().PutPolicy("test", `
path "pki/*" {
capabilities = ["update"]
}`)
if err != nil {
t.Fatal(err)
}
// Enable userpass auth method.
if err := client.Sys().EnableAuth("userpass", "userpass", ""); err != nil {
t.Fatal(err)
}
// Configure test role for userpass.
if _, err := client.Logical().Write("auth/userpass/users/userpassname", map[string]interface{}{
"password": "test",
"policies": "test",
}); err != nil {
t.Fatal(err)
}
// Login userpass for test role and keep client token.
secret, err := client.Logical().Write("auth/userpass/login/userpassname", map[string]interface{}{
"password": "test",
})
if err != nil || secret == nil {
t.Fatal(err)
}
userpassToken := secret.Auth.ClientToken
// Get auth accessor for identity template.
auths, err := client.Sys().ListAuth()
if err != nil {
t.Fatal(err)
}
userpassAccessor := auths["userpass/"].Accessor
// Mount PKI.
err = client.Sys().Mount("pki", &api.MountInput{
Type: "pki",
Config: api.MountConfigInput{
DefaultLeaseTTL: "16h",
MaxLeaseTTL: "60h",
},
})
if err != nil {
t.Fatal(err)
}
// Generate internal CA.
_, err = client.Logical().Write("pki/root/generate/internal", map[string]interface{}{
"ttl": "40h",
"common_name": "myvault.com",
})
if err != nil {
t.Fatal(err)
}
// Write role PKI.
_, err = client.Logical().Write("pki/roles/test", map[string]interface{}{
2022-01-27 18:06:34 +00:00
"allowed_uri_sans": []string{
"spiffe://domain/{{identity.entity.aliases." + userpassAccessor + ".name}}",
"spiffe://domain/{{identity.entity.aliases." + userpassAccessor + ".name}}/*", "spiffe://domain/foo",
},
"allowed_uri_sans_template": true,
"require_cn": false,
})
if err != nil {
t.Fatal(err)
}
// Issue certificate with identity templating
client.SetToken(userpassToken)
_, err = client.Logical().Write("pki/issue/test", map[string]interface{}{"uri_sans": "spiffe://domain/userpassname, spiffe://domain/foo"})
if err != nil {
t.Fatal(err)
}
// Issue certificate with identity templating and glob
client.SetToken(userpassToken)
_, err = client.Logical().Write("pki/issue/test", map[string]interface{}{"uri_sans": "spiffe://domain/userpassname/bar"})
if err != nil {
t.Fatal(err)
}
// Issue certificate with non-matching identity template parameter
client.SetToken(userpassToken)
_, err = client.Logical().Write("pki/issue/test", map[string]interface{}{"uri_sans": "spiffe://domain/unknownuser"})
if err == nil {
t.Fatal(err)
}
// Set allowed_uri_sans_template to false.
_, err = client.Logical().Write("pki/roles/test", map[string]interface{}{
"allowed_uri_sans_template": false,
})
if err != nil {
t.Fatal(err)
}
// Issue certificate with userpassToken.
_, err = client.Logical().Write("pki/issue/test", map[string]interface{}{"uri_sans": "spiffe://domain/users/userpassname"})
if err == nil {
t.Fatal("expected error")
}
}
func TestBackend_AllowedDomainsTemplate(t *testing.T) {
t.Parallel()
coreConfig := &vault.CoreConfig{
CredentialBackends: map[string]logical.Factory{
"userpass": userpass.Factory,
},
LogicalBackends: map[string]logical.Factory{
"pki": Factory,
},
}
cluster := vault.NewTestCluster(t, coreConfig, &vault.TestClusterOptions{
HandlerFunc: vaulthttp.Handler,
})
cluster.Start()
defer cluster.Cleanup()
client := cluster.Cores[0].Client
// Write test policy for userpass auth method.
err := client.Sys().PutPolicy("test", `
path "pki/*" {
capabilities = ["update"]
}`)
if err != nil {
t.Fatal(err)
}
// Enable userpass auth method.
if err := client.Sys().EnableAuth("userpass", "userpass", ""); err != nil {
t.Fatal(err)
}
// Configure test role for userpass.
if _, err := client.Logical().Write("auth/userpass/users/userpassname", map[string]interface{}{
"password": "test",
"policies": "test",
}); err != nil {
t.Fatal(err)
}
// Login userpass for test role and set client token
userpassAuth, err := auth.NewUserpassAuth("userpassname", &auth.Password{FromString: "test"})
if err != nil {
t.Fatal(err)
}
// Get auth accessor for identity template.
auths, err := client.Sys().ListAuth()
if err != nil {
t.Fatal(err)
}
userpassAccessor := auths["userpass/"].Accessor
// Mount PKI.
err = client.Sys().Mount("pki", &api.MountInput{
Type: "pki",
Config: api.MountConfigInput{
DefaultLeaseTTL: "16h",
MaxLeaseTTL: "60h",
},
})
if err != nil {
t.Fatal(err)
}
// Generate internal CA.
_, err = client.Logical().Write("pki/root/generate/internal", map[string]interface{}{
"ttl": "40h",
"common_name": "myvault.com",
})
if err != nil {
t.Fatal(err)
}
// Write role PKI.
_, err = client.Logical().Write("pki/roles/test", map[string]interface{}{
"allowed_domains": []string{
"foobar.com", "zipzap.com", "{{identity.entity.aliases." + userpassAccessor + ".name}}",
"foo.{{identity.entity.aliases." + userpassAccessor + ".name}}.example.com",
},
"allowed_domains_template": true,
"allow_bare_domains": true,
})
if err != nil {
t.Fatal(err)
}
// Issue certificate with userpassToken.
secret, err := client.Auth().Login(context.TODO(), userpassAuth)
if err != nil {
t.Fatal(err)
}
if err != nil || secret == nil {
t.Fatal(err)
}
_, err = client.Logical().Write("pki/issue/test", map[string]interface{}{"common_name": "userpassname"})
if err != nil {
t.Fatal(err)
}
// Issue certificate for foobar.com to verify allowed_domain_template doesn't break plain domains.
_, err = client.Logical().Write("pki/issue/test", map[string]interface{}{"common_name": "foobar.com"})
if err != nil {
t.Fatal(err)
}
// Issue certificate for unknown userpassname.
_, err = client.Logical().Write("pki/issue/test", map[string]interface{}{"common_name": "unknownuserpassname"})
if err == nil {
t.Fatal("expected error")
}
// Issue certificate for foo.userpassname.domain.
_, err = client.Logical().Write("pki/issue/test", map[string]interface{}{"common_name": "foo.userpassname.example.com"})
if err != nil {
t.Fatal("expected error")
}
// Set allowed_domains_template to false.
_, err = client.Logical().Write("pki/roles/test", map[string]interface{}{
"allowed_domains_template": false,
})
if err != nil {
t.Fatal(err)
}
// Issue certificate with userpassToken.
_, err = client.Logical().Write("pki/issue/test", map[string]interface{}{"common_name": "userpassname"})
if err == nil {
t.Fatal("expected error")
}
}
func TestReadWriteDeleteRoles(t *testing.T) {
t.Parallel()
ctx := context.Background()
coreConfig := &vault.CoreConfig{
CredentialBackends: map[string]logical.Factory{
"userpass": userpass.Factory,
},
LogicalBackends: map[string]logical.Factory{
"pki": Factory,
},
}
cluster := vault.NewTestCluster(t, coreConfig, &vault.TestClusterOptions{
HandlerFunc: vaulthttp.Handler,
})
cluster.Start()
defer cluster.Cleanup()
client := cluster.Cores[0].Client
// Mount PKI.
err := client.Sys().MountWithContext(ctx, "pki", &api.MountInput{
Type: "pki",
Config: api.MountConfigInput{
DefaultLeaseTTL: "16h",
MaxLeaseTTL: "60h",
},
})
if err != nil {
t.Fatal(err)
}
resp, err := client.Logical().ReadWithContext(ctx, "pki/roles/test")
if err != nil {
t.Fatal(err)
}
if resp != nil {
t.Fatalf("response should have been emtpy but was:\n%#v", resp)
}
// Write role PKI.
_, err = client.Logical().WriteWithContext(ctx, "pki/roles/test", map[string]interface{}{})
if err != nil {
t.Fatal(err)
}
// Read the role.
resp, err = client.Logical().ReadWithContext(ctx, "pki/roles/test")
if err != nil {
t.Fatal(err)
}
if resp.Data == nil {
t.Fatal("default data within response was nil when it should have contained data")
}
// Validate that we have not changed any defaults unknowingly
expectedData := map[string]interface{}{
"key_type": "rsa",
"use_csr_sans": true,
"client_flag": true,
"allowed_serial_numbers": []interface{}{},
"generate_lease": false,
"signature_bits": json.Number("256"),
Add PSS support to PKI Secrets Engine (#16519) * Add PSS signature support to Vault PKI engine Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use issuer's RevocationSigAlg for CRL signing We introduce a new parameter on issuers, revocation_signature_algorithm to control the signature algorithm used during CRL signing. This is because the SignatureAlgorithm value from the certificate itself is incorrect for this purpose: a RSA root could sign an ECDSA intermediate with say, SHA256WithRSA, but when the intermediate goes to sign a CRL, it must use ECDSAWithSHA256 or equivalent instead of SHA256WithRSA. When coupled with support for PSS-only keys, allowing the user to set the signature algorithm value as desired seems like the best approach. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add use_pss, revocation_signature_algorithm docs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add PSS to signature role issuance test matrix Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow roots to self-identify revocation alg When using PSS support with a managed key, sometimes the underlying device will not support PKCS#1v1.5 signatures. This results in CRL building failing, unless we update the entry's signature algorithm prior to building the CRL for the new root. With a RSA-type key and use_pss=true, we use the signature bits value to decide which hash function to use for PSS support. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add clearer error message on failed import When CRL building fails during cert/key import, due to PSS failures, give a better indication to the user that import succeeded its just CRL building that failed. This tells them the parameter to adjust on the issuer and warns that CRL building will fail until this is fixed. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add case insensitive SigAlgo matching Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert UsePSS back to regular bool Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor PSS->certTemplate into helper function Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Proper string output on rev_sig_alg display Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Copy root's SignatureAlgorithm for CRL building Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-03 16:42:24 +00:00
"use_pss": false,
"allowed_domains": []interface{}{},
"allowed_uri_sans_template": false,
"enforce_hostnames": true,
"policy_identifiers": []interface{}{},
"require_cn": true,
"allowed_domains_template": false,
"allow_token_displayname": false,
"country": []interface{}{},
"not_after": "",
"postal_code": []interface{}{},
"use_csr_common_name": true,
"allow_localhost": true,
"allow_subdomains": false,
"allow_wildcard_certificates": true,
"allowed_other_sans": []interface{}{},
"allowed_uri_sans": []interface{}{},
"basic_constraints_valid_for_non_ca": false,
"key_usage": []interface{}{"DigitalSignature", "KeyAgreement", "KeyEncipherment"},
"not_before_duration": json.Number("30"),
"allow_glob_domains": false,
"ttl": json.Number("0"),
"ou": []interface{}{},
"email_protection_flag": false,
"locality": []interface{}{},
"server_flag": true,
"allow_bare_domains": false,
"allow_ip_sans": true,
"ext_key_usage_oids": []interface{}{},
"allow_any_name": false,
"ext_key_usage": []interface{}{},
"key_bits": json.Number("2048"),
"max_ttl": json.Number("0"),
"no_store": false,
"organization": []interface{}{},
"province": []interface{}{},
"street_address": []interface{}{},
"code_signing_flag": false,
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"issuer_ref": "default",
"cn_validations": []interface{}{"email", "hostname"},
}
if diff := deep.Equal(expectedData, resp.Data); len(diff) > 0 {
t.Fatalf("pki role default values have changed, diff: %v", diff)
}
_, err = client.Logical().DeleteWithContext(ctx, "pki/roles/test")
if err != nil {
t.Fatal(err)
}
resp, err = client.Logical().ReadWithContext(ctx, "pki/roles/test")
if err != nil {
t.Fatal(err)
}
if resp != nil {
t.Fatalf("response should have been empty but was:\n%#v", resp)
}
}
2018-02-20 05:03:45 +00:00
func setCerts() {
cak, err := ecdsa.GenerateKey(elliptic.P256(), rand.Reader)
if err != nil {
panic(err)
}
marshaledKey, err := x509.MarshalECPrivateKey(cak)
if err != nil {
panic(err)
}
keyPEMBlock := &pem.Block{
Type: "EC PRIVATE KEY",
Bytes: marshaledKey,
}
ecCAKey = strings.TrimSpace(string(pem.EncodeToMemory(keyPEMBlock)))
2018-02-20 05:03:45 +00:00
if err != nil {
panic(err)
}
subjKeyID, err := certutil.GetSubjKeyID(cak)
if err != nil {
panic(err)
}
caCertTemplate := &x509.Certificate{
Subject: pkix.Name{
CommonName: "root.localhost",
},
SubjectKeyId: subjKeyID,
DNSNames: []string{"root.localhost"},
KeyUsage: x509.KeyUsage(x509.KeyUsageCertSign | x509.KeyUsageCRLSign),
SerialNumber: big.NewInt(mathrand.Int63()),
NotAfter: time.Now().Add(262980 * time.Hour),
BasicConstraintsValid: true,
2018-11-07 21:52:01 +00:00
IsCA: true,
2018-02-20 05:03:45 +00:00
}
caBytes, err := x509.CreateCertificate(rand.Reader, caCertTemplate, caCertTemplate, cak.Public(), cak)
if err != nil {
panic(err)
}
caCertPEMBlock := &pem.Block{
Type: "CERTIFICATE",
Bytes: caBytes,
}
ecCACert = strings.TrimSpace(string(pem.EncodeToMemory(caCertPEMBlock)))
2018-02-20 05:03:45 +00:00
rak, err := rsa.GenerateKey(rand.Reader, 2048)
if err != nil {
panic(err)
}
marshaledKey = x509.MarshalPKCS1PrivateKey(rak)
keyPEMBlock = &pem.Block{
Type: "RSA PRIVATE KEY",
Bytes: marshaledKey,
}
rsaCAKey = strings.TrimSpace(string(pem.EncodeToMemory(keyPEMBlock)))
2018-02-20 05:03:45 +00:00
if err != nil {
panic(err)
}
_, err = certutil.GetSubjKeyID(rak)
2018-02-20 05:03:45 +00:00
if err != nil {
panic(err)
}
caBytes, err = x509.CreateCertificate(rand.Reader, caCertTemplate, caCertTemplate, rak.Public(), rak)
if err != nil {
panic(err)
}
caCertPEMBlock = &pem.Block{
Type: "CERTIFICATE",
Bytes: caBytes,
}
rsaCACert = strings.TrimSpace(string(pem.EncodeToMemory(caCertPEMBlock)))
_, edk, err := ed25519.GenerateKey(rand.Reader)
if err != nil {
panic(err)
}
marshaledKey, err = x509.MarshalPKCS8PrivateKey(edk)
if err != nil {
panic(err)
}
keyPEMBlock = &pem.Block{
Type: "PRIVATE KEY",
Bytes: marshaledKey,
}
edCAKey = strings.TrimSpace(string(pem.EncodeToMemory(keyPEMBlock)))
if err != nil {
panic(err)
}
_, err = certutil.GetSubjKeyID(edk)
if err != nil {
panic(err)
}
caBytes, err = x509.CreateCertificate(rand.Reader, caCertTemplate, caCertTemplate, edk.Public(), edk)
if err != nil {
panic(err)
}
caCertPEMBlock = &pem.Block{
Type: "CERTIFICATE",
Bytes: caBytes,
}
edCACert = strings.TrimSpace(string(pem.EncodeToMemory(caCertPEMBlock)))
2018-02-20 05:03:45 +00:00
}
func TestBackend_RevokePlusTidy_Intermediate(t *testing.T) {
// Use a ridiculously long time to minimize the chance
// that we have to deal with more than one interval.
// InMemSink rounds down to an interval boundary rather than
// starting one at the time of initialization.
//
// This test is not parallelizable.
inmemSink := metrics.NewInmemSink(
1000000*time.Hour,
2000000*time.Hour)
metricsConf := metrics.DefaultConfig("")
metricsConf.EnableHostname = false
metricsConf.EnableHostnameLabel = false
metricsConf.EnableServiceLabel = false
metricsConf.EnableTypePrefix = false
_, err := metrics.NewGlobal(metricsConf, inmemSink)
if err != nil {
t.Fatal(err)
}
// Enable PKI secret engine
coreConfig := &vault.CoreConfig{
LogicalBackends: map[string]logical.Factory{
"pki": Factory,
},
}
cluster := vault.NewTestCluster(t, coreConfig, &vault.TestClusterOptions{
HandlerFunc: vaulthttp.Handler,
})
cluster.Start()
defer cluster.Cleanup()
cores := cluster.Cores
vault.TestWaitActive(t, cores[0].Core)
client := cores[0].Client
// Mount /pki as a root CA
err = client.Sys().Mount("pki", &api.MountInput{
Type: "pki",
Config: api.MountConfigInput{
DefaultLeaseTTL: "16h",
MaxLeaseTTL: "32h",
},
})
if err != nil {
t.Fatal(err)
}
// Check the metrics initialized in order to calculate backendUUID for /pki
// BackendUUID not consistent during tests with UUID from /sys/mounts/pki
metricsSuffix := "total_certificates_stored"
backendUUID := ""
mostRecentInterval := inmemSink.Data()[len(inmemSink.Data())-1]
for _, existingGauge := range mostRecentInterval.Gauges {
if strings.HasSuffix(existingGauge.Name, metricsSuffix) {
expandedGaugeName := existingGauge.Name
backendUUID = strings.Split(expandedGaugeName, ".")[2]
break
}
}
if backendUUID == "" {
t.Fatalf("No Gauge Found ending with %s", metricsSuffix)
}
// Set the cluster's certificate as the root CA in /pki
pemBundleRootCA := string(cluster.CACertPEM) + string(cluster.CAKeyPEM)
_, err = client.Logical().Write("pki/config/ca", map[string]interface{}{
"pem_bundle": pemBundleRootCA,
})
if err != nil {
t.Fatal(err)
}
// Mount /pki2 to operate as an intermediate CA
err = client.Sys().Mount("pki2", &api.MountInput{
Type: "pki",
Config: api.MountConfigInput{
DefaultLeaseTTL: "16h",
MaxLeaseTTL: "32h",
},
})
if err != nil {
t.Fatal(err)
}
// Create a CSR for the intermediate CA
secret, err := client.Logical().Write("pki2/intermediate/generate/internal", nil)
if err != nil {
t.Fatal(err)
}
intermediateCSR := secret.Data["csr"].(string)
// Sign the intermediate CSR using /pki
secret, err = client.Logical().Write("pki/root/sign-intermediate", map[string]interface{}{
"permitted_dns_domains": ".myvault.com",
"csr": intermediateCSR,
"ttl": "10s",
})
if err != nil {
t.Fatal(err)
}
intermediateCertSerial := secret.Data["serial_number"].(string)
intermediateCASerialColon := strings.ReplaceAll(strings.ToLower(intermediateCertSerial), ":", "-")
// Get the intermediate cert after signing
secret, err = client.Logical().Read("pki/cert/" + intermediateCASerialColon)
if err != nil {
t.Fatal(err)
}
if secret == nil || len(secret.Data) == 0 || len(secret.Data["certificate"].(string)) == 0 {
t.Fatal("expected certificate information from read operation")
}
// Issue a revoke on on /pki
_, err = client.Logical().Write("pki/revoke", map[string]interface{}{
"serial_number": intermediateCertSerial,
})
if err != nil {
t.Fatal(err)
}
// Check the cert-count metrics
expectedCertCountGaugeMetrics := map[string]float32{
"secrets.pki." + backendUUID + ".total_revoked_certificates_stored": 1,
"secrets.pki." + backendUUID + ".total_certificates_stored": 1,
}
mostRecentInterval = inmemSink.Data()[len(inmemSink.Data())-1]
for gauge, value := range expectedCertCountGaugeMetrics {
if _, ok := mostRecentInterval.Gauges[gauge]; !ok {
t.Fatalf("Expected metrics to include a value for gauge %s", gauge)
}
if value != mostRecentInterval.Gauges[gauge].Value {
t.Fatalf("Expected value metric %s to be %f but got %f", gauge, value, mostRecentInterval.Gauges[gauge].Value)
}
}
// Revoke adds a fixed 2s buffer, so we sleep for a bit longer to ensure
// the revocation time is past the current time.
time.Sleep(3 * time.Second)
// Issue a tidy on /pki
_, err = client.Logical().Write("pki/tidy", map[string]interface{}{
"tidy_cert_store": true,
"tidy_revoked_certs": true,
"safety_buffer": "1s",
})
if err != nil {
t.Fatal(err)
}
// Sleep a bit to make sure we're past the safety buffer
time.Sleep(2 * time.Second)
// Get CRL and ensure the tidied cert is still in the list after the tidy
// operation since it's not past the NotAfter (ttl) value yet.
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
crl := getParsedCrl(t, client, "pki")
revokedCerts := crl.TBSCertList.RevokedCertificates
if len(revokedCerts) == 0 {
t.Fatal("expected CRL to be non-empty")
}
sn := certutil.GetHexFormatted(revokedCerts[0].SerialNumber.Bytes(), ":")
if sn != intermediateCertSerial {
t.Fatalf("expected: %v, got: %v", intermediateCertSerial, sn)
}
// Wait for cert to expire
time.Sleep(10 * time.Second)
// Issue a tidy on /pki
_, err = client.Logical().Write("pki/tidy", map[string]interface{}{
"tidy_cert_store": true,
"tidy_revoked_certs": true,
"safety_buffer": "1s",
})
if err != nil {
t.Fatal(err)
}
// Sleep a bit to make sure we're past the safety buffer
time.Sleep(2 * time.Second)
// Issue a tidy-status on /pki
{
tidyStatus, err := client.Logical().Read("pki/tidy-status")
if err != nil {
t.Fatal(err)
}
expectedData := map[string]interface{}{
Let PKI tidy associate revoked certs with their issuers (#16871) * Refactor tidy steps into two separate helpers This refactors the tidy go routine into two separate helpers, making it clear where the boundaries of each are: variables are passed into these method and concerns are separated. As more operations are rolled into tidy, we can continue adding more helpers as appropriate. Additionally, as we move to make auto-tidy occur, we can use these as points to hook into periodic tidying. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor revInfo checking to helper This allows us to validate whether or not a revInfo entry contains a presently valid issuer, from the existing mapping. Coupled with the changeset to identify the issuer on revocation, we can begin adding capabilities to tidy to update this association, decreasing CRL build time and increasing the performance of OCSP. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor issuer fetching for revocation purposes Revocation needs to gracefully handle using the old legacy cert bundle, so fetching issuers (and parsing them) needs to be done slightly differently than other places. Refactor this from revokeCert into a common helper that can be used by tidy. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow tidy to associate revoked certs, issuers When revoking a certificate, we need to associate the issuer that signed its certificate back to the revInfo entry. Historically this was performed during CRL building (and still remains so), but when running without CRL building and with only OCSP, performance will degrade as the issuer needs to be found each time. Instead, allow the tidy operation to take over this role, allowing us to increase the performance of OCSP and CRL in this scenario, by decoupling issuer identification from CRL building in the ideal case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for tidy updates Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation on new tidy parameter, metrics Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor tidy config into shared struct Finish adding metrics, status messages about new tidy operation. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-26 17:13:45 +00:00
"safety_buffer": json.Number("1"),
Add automatic tidy of expired issuers (#17823) * Add automatic tidy of expired issuers To aid PKI users like Consul, which periodically rotate intermediates, and provided a little more consistency with older versions of Vault which would silently (and dangerously!) replace the configured CA on root/intermediate generation, we introduce an automatic tidy of expired issuers. This includes a longer safety buffer (1 year) and logging of the relevant issuer information prior to deletion (certificate contents, key ID, and issuer ID/name) to allow admins to recover this value if desired, or perform further cleanup of keys. From my PoV, removal of the issuer is thus a relatively safe operation compared to keys (which I do not feel comfortable removing) as they can always be re-imported if desired. Additionally, this is an opt-in tidy operation, not enabled by default. Lastly, most major performance penalties comes with lots of issuers within the mount, not as much large numbers of keys (as only new issuer creation/import operations are affected, unlike LIST /issuers which is a public, unauthenticated endpoint). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add test for tidy Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add docs on tidy of issuers Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Restructure logging Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing fields to expected tidy output Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-11-10 15:53:26 +00:00
"issuer_safety_buffer": json.Number("31536000"),
Let PKI tidy associate revoked certs with their issuers (#16871) * Refactor tidy steps into two separate helpers This refactors the tidy go routine into two separate helpers, making it clear where the boundaries of each are: variables are passed into these method and concerns are separated. As more operations are rolled into tidy, we can continue adding more helpers as appropriate. Additionally, as we move to make auto-tidy occur, we can use these as points to hook into periodic tidying. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor revInfo checking to helper This allows us to validate whether or not a revInfo entry contains a presently valid issuer, from the existing mapping. Coupled with the changeset to identify the issuer on revocation, we can begin adding capabilities to tidy to update this association, decreasing CRL build time and increasing the performance of OCSP. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor issuer fetching for revocation purposes Revocation needs to gracefully handle using the old legacy cert bundle, so fetching issuers (and parsing them) needs to be done slightly differently than other places. Refactor this from revokeCert into a common helper that can be used by tidy. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow tidy to associate revoked certs, issuers When revoking a certificate, we need to associate the issuer that signed its certificate back to the revInfo entry. Historically this was performed during CRL building (and still remains so), but when running without CRL building and with only OCSP, performance will degrade as the issuer needs to be found each time. Instead, allow the tidy operation to take over this role, allowing us to increase the performance of OCSP and CRL in this scenario, by decoupling issuer identification from CRL building in the ideal case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for tidy updates Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation on new tidy parameter, metrics Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor tidy config into shared struct Finish adding metrics, status messages about new tidy operation. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-26 17:13:45 +00:00
"tidy_cert_store": true,
"tidy_revoked_certs": true,
"tidy_revoked_cert_issuer_associations": false,
Add automatic tidy of expired issuers (#17823) * Add automatic tidy of expired issuers To aid PKI users like Consul, which periodically rotate intermediates, and provided a little more consistency with older versions of Vault which would silently (and dangerously!) replace the configured CA on root/intermediate generation, we introduce an automatic tidy of expired issuers. This includes a longer safety buffer (1 year) and logging of the relevant issuer information prior to deletion (certificate contents, key ID, and issuer ID/name) to allow admins to recover this value if desired, or perform further cleanup of keys. From my PoV, removal of the issuer is thus a relatively safe operation compared to keys (which I do not feel comfortable removing) as they can always be re-imported if desired. Additionally, this is an opt-in tidy operation, not enabled by default. Lastly, most major performance penalties comes with lots of issuers within the mount, not as much large numbers of keys (as only new issuer creation/import operations are affected, unlike LIST /issuers which is a public, unauthenticated endpoint). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add test for tidy Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add docs on tidy of issuers Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Restructure logging Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing fields to expected tidy output Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-11-10 15:53:26 +00:00
"tidy_expired_issuers": false,
Allow tidy to backup legacy CA bundles (#18645) * Allow tidy to backup legacy CA bundles With the new tidy_move_legacy_ca_bundle option, we'll use tidy to move the legacy CA bundle from /config/ca_bundle to /config/ca_bundle.bak. This does two things: 1. Removes ca_bundle from the hot-path of initialization after initial migration has completed. Because this entry is seal wrapped, this may result in performance improvements. 2. Allows recovery of this value in the event of some other failure with migration. Notably, this cannot occur during migration in the unlikely (and largely unsupported) case that the operator immediately downgrades to Vault <1.11.x. Thus, we reuse issuer_safety_buffer; while potentially long, tidy can always be run manually with a shorter buffer (and only this flag) to manually move the bundle if necessary. In the event of needing to recover or undo this operation, it is sufficient to use sys/raw to read the backed up value and subsequently write it to its old path (/config/ca_bundle). The new entry remains seal wrapped, but otherwise isn't used within the code and so has better performance characteristics. Performing a fat deletion (DELETE /root) will again remove the backup like the old legacy bundle, preserving its wipe characteristics. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation about new tidy parameter Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for migration scenarios Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up time comparisons Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2023-01-11 17:12:53 +00:00
"tidy_move_legacy_ca_bundle": false,
Add ability to cancel PKI tidy operations, pause between tidying certs (#16958) * Allow tidy operations to be cancelled When tidy operations take a long time to execute (and especially when executing them automatically), having the ability to cancel them becomes useful to reduce strain on Vault clusters (and let them be rescheduled at a later time). To this end, we add the /tidy-cancel write endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing auto-tidy synopsis / description Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add a pause duration between tidying certificates By setting pause_duration, operators can have a little control over the resource utilization of a tidy operation. While the list of certificates remain in memory throughout the entire operation, a pause is added between processing certificates and the revocation lock is released. This allows other operations to occur during this gap and potentially allows the tidy operation to consume less resources per unit of time (due to the sleep -- though obviously consumes the same resources over the time of the operation). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for cancellation, pause Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add API docs on pause_duration, /tidy-cancel Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add lock releasing around tidy pause Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Reset cancel guard, return errors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-31 18:36:12 +00:00
"pause_duration": "0s",
Let PKI tidy associate revoked certs with their issuers (#16871) * Refactor tidy steps into two separate helpers This refactors the tidy go routine into two separate helpers, making it clear where the boundaries of each are: variables are passed into these method and concerns are separated. As more operations are rolled into tidy, we can continue adding more helpers as appropriate. Additionally, as we move to make auto-tidy occur, we can use these as points to hook into periodic tidying. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor revInfo checking to helper This allows us to validate whether or not a revInfo entry contains a presently valid issuer, from the existing mapping. Coupled with the changeset to identify the issuer on revocation, we can begin adding capabilities to tidy to update this association, decreasing CRL build time and increasing the performance of OCSP. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor issuer fetching for revocation purposes Revocation needs to gracefully handle using the old legacy cert bundle, so fetching issuers (and parsing them) needs to be done slightly differently than other places. Refactor this from revokeCert into a common helper that can be used by tidy. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow tidy to associate revoked certs, issuers When revoking a certificate, we need to associate the issuer that signed its certificate back to the revInfo entry. Historically this was performed during CRL building (and still remains so), but when running without CRL building and with only OCSP, performance will degrade as the issuer needs to be found each time. Instead, allow the tidy operation to take over this role, allowing us to increase the performance of OCSP and CRL in this scenario, by decoupling issuer identification from CRL building in the ideal case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for tidy updates Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation on new tidy parameter, metrics Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor tidy config into shared struct Finish adding metrics, status messages about new tidy operation. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-26 17:13:45 +00:00
"state": "Finished",
"error": nil,
"time_started": nil,
"time_finished": nil,
"message": nil,
"cert_store_deleted_count": json.Number("1"),
"revoked_cert_deleted_count": json.Number("1"),
"missing_issuer_cert_count": json.Number("0"),
"current_cert_store_count": json.Number("0"),
"current_revoked_cert_count": json.Number("0"),
"revocation_queue_deleted_count": json.Number("0"),
}
// Let's copy the times from the response so that we can use deep.Equal()
timeStarted, ok := tidyStatus.Data["time_started"]
if !ok || timeStarted == "" {
t.Fatal("Expected tidy status response to include a value for time_started")
}
expectedData["time_started"] = timeStarted
timeFinished, ok := tidyStatus.Data["time_finished"]
if !ok || timeFinished == "" {
t.Fatal("Expected tidy status response to include a value for time_finished")
}
expectedData["time_finished"] = timeFinished
if diff := deep.Equal(expectedData, tidyStatus.Data); diff != nil {
t.Fatal(diff)
}
}
// Check the tidy metrics
{
// Map of gauges to expected value
expectedGauges := map[string]float32{
"secrets.pki.tidy.cert_store_current_entry": 0,
"secrets.pki.tidy.cert_store_total_entries": 1,
"secrets.pki.tidy.revoked_cert_current_entry": 0,
"secrets.pki.tidy.revoked_cert_total_entries": 1,
"secrets.pki.tidy.start_time_epoch": 0,
"secrets.pki." + backendUUID + ".total_certificates_stored": 0,
"secrets.pki." + backendUUID + ".total_revoked_certificates_stored": 0,
"secrets.pki.tidy.cert_store_total_entries_remaining": 0,
"secrets.pki.tidy.revoked_cert_total_entries_remaining": 0,
}
// Map of counters to the sum of the metrics for that counter
expectedCounters := map[string]float64{
"secrets.pki.tidy.cert_store_deleted_count": 1,
"secrets.pki.tidy.revoked_cert_deleted_count": 1,
"secrets.pki.tidy.success": 2,
// Note that "secrets.pki.tidy.failure" won't be in the captured metrics
}
// If the metrics span more than one interval, skip the checks
intervals := inmemSink.Data()
if len(intervals) == 1 {
interval := inmemSink.Data()[0]
for gauge, value := range expectedGauges {
if _, ok := interval.Gauges[gauge]; !ok {
t.Fatalf("Expected metrics to include a value for gauge %s", gauge)
}
if value != interval.Gauges[gauge].Value {
t.Fatalf("Expected value metric %s to be %f but got %f", gauge, value, interval.Gauges[gauge].Value)
}
}
for counter, value := range expectedCounters {
if _, ok := interval.Counters[counter]; !ok {
t.Fatalf("Expected metrics to include a value for couter %s", counter)
}
if value != interval.Counters[counter].Sum {
t.Fatalf("Expected the sum of metric %s to be %f but got %f", counter, value, interval.Counters[counter].Sum)
}
}
tidyDuration, ok := interval.Samples["secrets.pki.tidy.duration"]
if !ok {
t.Fatal("Expected metrics to include a value for sample secrets.pki.tidy.duration")
}
if tidyDuration.Count <= 0 {
t.Fatalf("Expected metrics to have count > 0 for sample secrets.pki.tidy.duration, but got %d", tidyDuration.Count)
}
}
}
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
crl = getParsedCrl(t, client, "pki")
revokedCerts = crl.TBSCertList.RevokedCertificates
if len(revokedCerts) != 0 {
t.Fatal("expected CRL to be empty")
}
}
func TestBackend_Root_FullCAChain(t *testing.T) {
t.Parallel()
testCases := []struct {
testName string
keyType string
}{
{testName: "RSA", keyType: "rsa"},
{testName: "ED25519", keyType: "ed25519"},
{testName: "EC", keyType: "ec"},
}
for _, tc := range testCases {
tc := tc
t.Run(tc.testName, func(t *testing.T) {
runFullCAChainTest(t, tc.keyType)
})
}
}
func runFullCAChainTest(t *testing.T, keyType string) {
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
// Generate a root CA at /pki-root
b_root, s_root := CreateBackendWithStorage(t)
var err error
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err := CBWrite(b_root, s_root, "root/generate/exported", map[string]interface{}{
"common_name": "root myvault.com",
"key_type": keyType,
})
if err != nil {
t.Fatal(err)
}
if resp == nil {
t.Fatal("expected ca info")
}
rootData := resp.Data
rootCert := rootData["certificate"].(string)
// Validate that root's /cert/ca-chain now contains the certificate.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBRead(b_root, s_root, "cert/ca_chain")
if err != nil {
t.Fatal(err)
}
if resp == nil {
t.Fatal("expected intermediate chain information")
}
fullChain := resp.Data["ca_chain"].(string)
requireCertInCaChainString(t, fullChain, rootCert, "expected root cert within root cert/ca_chain")
// Make sure when we issue a leaf certificate we get the full chain back.
_, err = CBWrite(b_root, s_root, "roles/example", map[string]interface{}{
"allowed_domains": "example.com",
"allow_subdomains": "true",
"max_ttl": "1h",
})
require.NoError(t, err, "error setting up pki root role: %v", err)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b_root, s_root, "issue/example", map[string]interface{}{
"common_name": "test.example.com",
"ttl": "5m",
})
require.NoError(t, err, "error issuing certificate from pki root: %v", err)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
fullChainArray := resp.Data["ca_chain"].([]string)
requireCertInCaChainArray(t, fullChainArray, rootCert, "expected root cert within root issuance pki-root/issue/example")
// Now generate an intermediate at /pki-intermediate, signed by the root.
b_int, s_int := CreateBackendWithStorage(t)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b_int, s_int, "intermediate/generate/exported", map[string]interface{}{
"common_name": "intermediate myvault.com",
"key_type": keyType,
})
if err != nil {
t.Fatal(err)
}
if resp == nil {
t.Fatal("expected intermediate CSR info")
}
intermediateData := resp.Data
intermediateKey := intermediateData["private_key"].(string)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b_root, s_root, "root/sign-intermediate", map[string]interface{}{
"csr": intermediateData["csr"],
"format": "pem",
})
if err != nil {
t.Fatal(err)
}
if resp == nil {
t.Fatal("expected signed intermediate info")
}
intermediateSignedData := resp.Data
intermediateCert := intermediateSignedData["certificate"].(string)
rootCaCert := parseCert(t, rootCert)
intermediaryCaCert := parseCert(t, intermediateCert)
requireSignedBy(t, intermediaryCaCert, rootCaCert)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
intermediateCaChain := intermediateSignedData["ca_chain"].([]string)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
require.Equal(t, parseCert(t, intermediateCaChain[0]), intermediaryCaCert, "intermediate signed cert should have been part of ca_chain")
require.Equal(t, parseCert(t, intermediateCaChain[1]), rootCaCert, "root cert should have been part of ca_chain")
_, err = CBWrite(b_int, s_int, "intermediate/set-signed", map[string]interface{}{
"certificate": intermediateCert + "\n" + rootCert + "\n",
})
if err != nil {
t.Fatal(err)
}
// Validate that intermediate's ca_chain field now includes the full
// chain.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBRead(b_int, s_int, "cert/ca_chain")
if err != nil {
t.Fatal(err)
}
if resp == nil {
t.Fatal("expected intermediate chain information")
}
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
// Verify we have a proper CRL now
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
crl := getParsedCrlFromBackend(t, b_int, s_int, "crl")
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
require.Equal(t, 0, len(crl.TBSCertList.RevokedCertificates))
fullChain = resp.Data["ca_chain"].(string)
requireCertInCaChainString(t, fullChain, intermediateCert, "expected full chain to contain intermediate certificate from pki-intermediate/cert/ca_chain")
requireCertInCaChainString(t, fullChain, rootCert, "expected full chain to contain root certificate from pki-intermediate/cert/ca_chain")
// Make sure when we issue a leaf certificate we get the full chain back.
_, err = CBWrite(b_int, s_int, "roles/example", map[string]interface{}{
"allowed_domains": "example.com",
"allow_subdomains": "true",
"max_ttl": "1h",
})
require.NoError(t, err, "error setting up pki intermediate role: %v", err)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b_int, s_int, "issue/example", map[string]interface{}{
"common_name": "test.example.com",
"ttl": "5m",
})
require.NoError(t, err, "error issuing certificate from pki intermediate: %v", err)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
fullChainArray = resp.Data["ca_chain"].([]string)
requireCertInCaChainArray(t, fullChainArray, intermediateCert, "expected full chain to contain intermediate certificate from pki-intermediate/issue/example")
requireCertInCaChainArray(t, fullChainArray, rootCert, "expected full chain to contain root certificate from pki-intermediate/issue/example")
// Finally, import this signing cert chain into a new mount to ensure
// "external" CAs behave as expected.
b_ext, s_ext := CreateBackendWithStorage(t)
_, err = CBWrite(b_ext, s_ext, "config/ca", map[string]interface{}{
"pem_bundle": intermediateKey + "\n" + intermediateCert + "\n" + rootCert + "\n",
})
if err != nil {
t.Fatal(err)
}
// Validate the external chain information was loaded correctly.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBRead(b_ext, s_ext, "cert/ca_chain")
if err != nil {
t.Fatal(err)
}
if resp == nil {
t.Fatal("expected intermediate chain information")
}
fullChain = resp.Data["ca_chain"].(string)
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
if strings.Count(fullChain, intermediateCert) != 1 {
t.Fatalf("expected full chain to contain intermediate certificate; got %v occurrences", strings.Count(fullChain, intermediateCert))
}
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
if strings.Count(fullChain, rootCert) != 1 {
t.Fatalf("expected full chain to contain root certificate; got %v occurrences", strings.Count(fullChain, rootCert))
}
// Now issue a short-lived certificate from our pki-external.
_, err = CBWrite(b_ext, s_ext, "roles/example", map[string]interface{}{
"allowed_domains": "example.com",
"allow_subdomains": "true",
"max_ttl": "1h",
})
require.NoError(t, err, "error setting up pki role: %v", err)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b_ext, s_ext, "issue/example", map[string]interface{}{
"common_name": "test.example.com",
"ttl": "5m",
})
require.NoError(t, err, "error issuing certificate: %v", err)
require.NotNil(t, resp, "got nil response from issuing request")
issueCrtAsPem := resp.Data["certificate"].(string)
issuedCrt := parseCert(t, issueCrtAsPem)
// Verify that the certificates are signed by the intermediary CA key...
requireSignedBy(t, issuedCrt, intermediaryCaCert)
// Test that we can request that the root ca certificate not appear in the ca_chain field
resp, err = CBWrite(b_ext, s_ext, "issue/example", map[string]interface{}{
"common_name": "test.example.com",
"ttl": "5m",
"remove_roots_from_chain": "true",
})
requireSuccessNonNilResponse(t, resp, err, "error issuing certificate when removing self signed")
fullChain = strings.Join(resp.Data["ca_chain"].([]string), "\n")
if strings.Count(fullChain, intermediateCert) != 1 {
t.Fatalf("expected full chain to contain intermediate certificate; got %v occurrences", strings.Count(fullChain, intermediateCert))
}
if strings.Count(fullChain, rootCert) != 0 {
t.Fatalf("expected full chain to NOT contain root certificate; got %v occurrences", strings.Count(fullChain, rootCert))
}
}
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
func requireCertInCaChainArray(t *testing.T, chain []string, cert string, msgAndArgs ...interface{}) {
var fullChain string
for _, caCert := range chain {
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
fullChain = fullChain + "\n" + caCert
}
requireCertInCaChainString(t, fullChain, cert, msgAndArgs)
}
func requireCertInCaChainString(t *testing.T, chain string, cert string, msgAndArgs ...interface{}) {
count := strings.Count(chain, cert)
if count != 1 {
failMsg := fmt.Sprintf("Found %d occurrances of the cert in the provided chain", count)
require.FailNow(t, failMsg, msgAndArgs...)
}
}
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
type MultiBool int
const (
MFalse MultiBool = iota
MTrue MultiBool = iota
MAny MultiBool = iota
)
func (o MultiBool) ToValues() []bool {
if o == MTrue {
return []bool{true}
}
if o == MFalse {
return []bool{false}
}
if o == MAny {
return []bool{true, false}
}
return []bool{}
}
type IssuanceRegression struct {
Add role parameter to restrict issuance of wildcard certificates (#14238) * Add new AllowWildcardCertificate field to PKI role This field allows the PKI role to control whether or not issuance of wildcard certificates are allowed. We default (both on migration and new role creation) to the less secure true value for backwards compatibility with existing Vault versions. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sanitizedName to reducedName Per comment, this variable name was confusing during the reproduction and subsequent fix of the earlier vulnerability and associated bug report. Because the common name isn't necessarily _sanitized_ in any way (and indeed must be considered in relation to other parts or the whole), but portions of the entire name are removed, reducedName appears to make the most sense. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Enforce AllowWildcardCertificates during issuance This commit adds the bulk of correctly validating wildcard certificate Common Names during issuance according to RFC 6125 Section 6.4.3 semantics. As part of this, support for RFC 2818-conforming wildcard certificates (wherein there are almost no restrictions on issuance) has been removed. Note that this flag does take precedence over AllowAnyName, giving a little more safety in wildcard issuance in this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test cases to conform with RFC 6125 Test cases 19, 70+71, and 83+84 didn't conform with the RFC 6125, and so should've been rejected under strict conformance. For 70+71 and 83+84, we previously conditioned around the value of AllowSubdomains (allowing issuance when true), but they likely should've been rejected either way. Additionally, update the notes about globs matching wildcard certificates to notate this is indeed the case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Check AllowWildcardCertifciates in issuance tests This allows for regression tests to cover the new AllowWildcardCertificate conditional. We add additional test cases ensuring that wildcard issuance is properly forbidden in all relevant scenarios, while allowing the existing test cases to validate that wildcard status doesn't affect non-wildcard certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add Wildcard allowance during signing operations When using sign-verbatim, sign-intermediate, or getting certificate generation parameters, set AllowWildcardCertificates to mirror existing policies. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-02-24 13:41:56 +00:00
AllowedDomains []string
AllowBareDomains MultiBool
AllowGlobDomains MultiBool
AllowSubdomains MultiBool
AllowLocalhost MultiBool
AllowWildcardCertificates MultiBool
CNValidations []string
Add role parameter to restrict issuance of wildcard certificates (#14238) * Add new AllowWildcardCertificate field to PKI role This field allows the PKI role to control whether or not issuance of wildcard certificates are allowed. We default (both on migration and new role creation) to the less secure true value for backwards compatibility with existing Vault versions. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sanitizedName to reducedName Per comment, this variable name was confusing during the reproduction and subsequent fix of the earlier vulnerability and associated bug report. Because the common name isn't necessarily _sanitized_ in any way (and indeed must be considered in relation to other parts or the whole), but portions of the entire name are removed, reducedName appears to make the most sense. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Enforce AllowWildcardCertificates during issuance This commit adds the bulk of correctly validating wildcard certificate Common Names during issuance according to RFC 6125 Section 6.4.3 semantics. As part of this, support for RFC 2818-conforming wildcard certificates (wherein there are almost no restrictions on issuance) has been removed. Note that this flag does take precedence over AllowAnyName, giving a little more safety in wildcard issuance in this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test cases to conform with RFC 6125 Test cases 19, 70+71, and 83+84 didn't conform with the RFC 6125, and so should've been rejected under strict conformance. For 70+71 and 83+84, we previously conditioned around the value of AllowSubdomains (allowing issuance when true), but they likely should've been rejected either way. Additionally, update the notes about globs matching wildcard certificates to notate this is indeed the case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Check AllowWildcardCertifciates in issuance tests This allows for regression tests to cover the new AllowWildcardCertificate conditional. We add additional test cases ensuring that wildcard issuance is properly forbidden in all relevant scenarios, while allowing the existing test cases to validate that wildcard status doesn't affect non-wildcard certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add Wildcard allowance during signing operations When using sign-verbatim, sign-intermediate, or getting certificate generation parameters, set AllowWildcardCertificates to mirror existing policies. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-02-24 13:41:56 +00:00
CommonName string
Issued bool
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
}
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
func RoleIssuanceRegressionHelper(t *testing.T, b *backend, s logical.Storage, index int, test IssuanceRegression) int {
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
tested := 0
for _, AllowBareDomains := range test.AllowBareDomains.ToValues() {
for _, AllowGlobDomains := range test.AllowGlobDomains.ToValues() {
for _, AllowSubdomains := range test.AllowSubdomains.ToValues() {
for _, AllowLocalhost := range test.AllowLocalhost.ToValues() {
Add role parameter to restrict issuance of wildcard certificates (#14238) * Add new AllowWildcardCertificate field to PKI role This field allows the PKI role to control whether or not issuance of wildcard certificates are allowed. We default (both on migration and new role creation) to the less secure true value for backwards compatibility with existing Vault versions. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sanitizedName to reducedName Per comment, this variable name was confusing during the reproduction and subsequent fix of the earlier vulnerability and associated bug report. Because the common name isn't necessarily _sanitized_ in any way (and indeed must be considered in relation to other parts or the whole), but portions of the entire name are removed, reducedName appears to make the most sense. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Enforce AllowWildcardCertificates during issuance This commit adds the bulk of correctly validating wildcard certificate Common Names during issuance according to RFC 6125 Section 6.4.3 semantics. As part of this, support for RFC 2818-conforming wildcard certificates (wherein there are almost no restrictions on issuance) has been removed. Note that this flag does take precedence over AllowAnyName, giving a little more safety in wildcard issuance in this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test cases to conform with RFC 6125 Test cases 19, 70+71, and 83+84 didn't conform with the RFC 6125, and so should've been rejected under strict conformance. For 70+71 and 83+84, we previously conditioned around the value of AllowSubdomains (allowing issuance when true), but they likely should've been rejected either way. Additionally, update the notes about globs matching wildcard certificates to notate this is indeed the case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Check AllowWildcardCertifciates in issuance tests This allows for regression tests to cover the new AllowWildcardCertificate conditional. We add additional test cases ensuring that wildcard issuance is properly forbidden in all relevant scenarios, while allowing the existing test cases to validate that wildcard status doesn't affect non-wildcard certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add Wildcard allowance during signing operations When using sign-verbatim, sign-intermediate, or getting certificate generation parameters, set AllowWildcardCertificates to mirror existing policies. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-02-24 13:41:56 +00:00
for _, AllowWildcardCertificates := range test.AllowWildcardCertificates.ToValues() {
role := fmt.Sprintf("issuance-regression-%d-bare-%v-glob-%v-subdomains-%v-localhost-%v-wildcard-%v", index, AllowBareDomains, AllowGlobDomains, AllowSubdomains, AllowLocalhost, AllowWildcardCertificates)
_, err := CBWrite(b, s, "roles/"+role, map[string]interface{}{
Add role parameter to restrict issuance of wildcard certificates (#14238) * Add new AllowWildcardCertificate field to PKI role This field allows the PKI role to control whether or not issuance of wildcard certificates are allowed. We default (both on migration and new role creation) to the less secure true value for backwards compatibility with existing Vault versions. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sanitizedName to reducedName Per comment, this variable name was confusing during the reproduction and subsequent fix of the earlier vulnerability and associated bug report. Because the common name isn't necessarily _sanitized_ in any way (and indeed must be considered in relation to other parts or the whole), but portions of the entire name are removed, reducedName appears to make the most sense. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Enforce AllowWildcardCertificates during issuance This commit adds the bulk of correctly validating wildcard certificate Common Names during issuance according to RFC 6125 Section 6.4.3 semantics. As part of this, support for RFC 2818-conforming wildcard certificates (wherein there are almost no restrictions on issuance) has been removed. Note that this flag does take precedence over AllowAnyName, giving a little more safety in wildcard issuance in this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test cases to conform with RFC 6125 Test cases 19, 70+71, and 83+84 didn't conform with the RFC 6125, and so should've been rejected under strict conformance. For 70+71 and 83+84, we previously conditioned around the value of AllowSubdomains (allowing issuance when true), but they likely should've been rejected either way. Additionally, update the notes about globs matching wildcard certificates to notate this is indeed the case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Check AllowWildcardCertifciates in issuance tests This allows for regression tests to cover the new AllowWildcardCertificate conditional. We add additional test cases ensuring that wildcard issuance is properly forbidden in all relevant scenarios, while allowing the existing test cases to validate that wildcard status doesn't affect non-wildcard certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add Wildcard allowance during signing operations When using sign-verbatim, sign-intermediate, or getting certificate generation parameters, set AllowWildcardCertificates to mirror existing policies. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-02-24 13:41:56 +00:00
"allowed_domains": test.AllowedDomains,
"allow_bare_domains": AllowBareDomains,
"allow_glob_domains": AllowGlobDomains,
"allow_subdomains": AllowSubdomains,
"allow_localhost": AllowLocalhost,
"allow_wildcard_certificates": AllowWildcardCertificates,
"cn_validations": test.CNValidations,
Add role parameter to restrict issuance of wildcard certificates (#14238) * Add new AllowWildcardCertificate field to PKI role This field allows the PKI role to control whether or not issuance of wildcard certificates are allowed. We default (both on migration and new role creation) to the less secure true value for backwards compatibility with existing Vault versions. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sanitizedName to reducedName Per comment, this variable name was confusing during the reproduction and subsequent fix of the earlier vulnerability and associated bug report. Because the common name isn't necessarily _sanitized_ in any way (and indeed must be considered in relation to other parts or the whole), but portions of the entire name are removed, reducedName appears to make the most sense. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Enforce AllowWildcardCertificates during issuance This commit adds the bulk of correctly validating wildcard certificate Common Names during issuance according to RFC 6125 Section 6.4.3 semantics. As part of this, support for RFC 2818-conforming wildcard certificates (wherein there are almost no restrictions on issuance) has been removed. Note that this flag does take precedence over AllowAnyName, giving a little more safety in wildcard issuance in this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test cases to conform with RFC 6125 Test cases 19, 70+71, and 83+84 didn't conform with the RFC 6125, and so should've been rejected under strict conformance. For 70+71 and 83+84, we previously conditioned around the value of AllowSubdomains (allowing issuance when true), but they likely should've been rejected either way. Additionally, update the notes about globs matching wildcard certificates to notate this is indeed the case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Check AllowWildcardCertifciates in issuance tests This allows for regression tests to cover the new AllowWildcardCertificate conditional. We add additional test cases ensuring that wildcard issuance is properly forbidden in all relevant scenarios, while allowing the existing test cases to validate that wildcard status doesn't affect non-wildcard certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add Wildcard allowance during signing operations When using sign-verbatim, sign-intermediate, or getting certificate generation parameters, set AllowWildcardCertificates to mirror existing policies. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-02-24 13:41:56 +00:00
// TODO: test across this vector as well. Currently certain wildcard
// matching is broken with it enabled (such as x*x.foo).
"enforce_hostnames": false,
"key_type": "ec",
"key_bits": 256,
"no_store": true,
// With the CN Validations field, ensure we prevent CN from appearing
// in SANs.
Add role parameter to restrict issuance of wildcard certificates (#14238) * Add new AllowWildcardCertificate field to PKI role This field allows the PKI role to control whether or not issuance of wildcard certificates are allowed. We default (both on migration and new role creation) to the less secure true value for backwards compatibility with existing Vault versions. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sanitizedName to reducedName Per comment, this variable name was confusing during the reproduction and subsequent fix of the earlier vulnerability and associated bug report. Because the common name isn't necessarily _sanitized_ in any way (and indeed must be considered in relation to other parts or the whole), but portions of the entire name are removed, reducedName appears to make the most sense. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Enforce AllowWildcardCertificates during issuance This commit adds the bulk of correctly validating wildcard certificate Common Names during issuance according to RFC 6125 Section 6.4.3 semantics. As part of this, support for RFC 2818-conforming wildcard certificates (wherein there are almost no restrictions on issuance) has been removed. Note that this flag does take precedence over AllowAnyName, giving a little more safety in wildcard issuance in this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test cases to conform with RFC 6125 Test cases 19, 70+71, and 83+84 didn't conform with the RFC 6125, and so should've been rejected under strict conformance. For 70+71 and 83+84, we previously conditioned around the value of AllowSubdomains (allowing issuance when true), but they likely should've been rejected either way. Additionally, update the notes about globs matching wildcard certificates to notate this is indeed the case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Check AllowWildcardCertifciates in issuance tests This allows for regression tests to cover the new AllowWildcardCertificate conditional. We add additional test cases ensuring that wildcard issuance is properly forbidden in all relevant scenarios, while allowing the existing test cases to validate that wildcard status doesn't affect non-wildcard certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add Wildcard allowance during signing operations When using sign-verbatim, sign-intermediate, or getting certificate generation parameters, set AllowWildcardCertificates to mirror existing policies. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-02-24 13:41:56 +00:00
})
if err != nil {
t.Fatal(err)
}
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
resp, err := CBWrite(b, s, "issue/"+role, map[string]interface{}{
"common_name": test.CommonName,
"exclude_cn_from_sans": true,
Add role parameter to restrict issuance of wildcard certificates (#14238) * Add new AllowWildcardCertificate field to PKI role This field allows the PKI role to control whether or not issuance of wildcard certificates are allowed. We default (both on migration and new role creation) to the less secure true value for backwards compatibility with existing Vault versions. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sanitizedName to reducedName Per comment, this variable name was confusing during the reproduction and subsequent fix of the earlier vulnerability and associated bug report. Because the common name isn't necessarily _sanitized_ in any way (and indeed must be considered in relation to other parts or the whole), but portions of the entire name are removed, reducedName appears to make the most sense. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Enforce AllowWildcardCertificates during issuance This commit adds the bulk of correctly validating wildcard certificate Common Names during issuance according to RFC 6125 Section 6.4.3 semantics. As part of this, support for RFC 2818-conforming wildcard certificates (wherein there are almost no restrictions on issuance) has been removed. Note that this flag does take precedence over AllowAnyName, giving a little more safety in wildcard issuance in this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test cases to conform with RFC 6125 Test cases 19, 70+71, and 83+84 didn't conform with the RFC 6125, and so should've been rejected under strict conformance. For 70+71 and 83+84, we previously conditioned around the value of AllowSubdomains (allowing issuance when true), but they likely should've been rejected either way. Additionally, update the notes about globs matching wildcard certificates to notate this is indeed the case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Check AllowWildcardCertifciates in issuance tests This allows for regression tests to cover the new AllowWildcardCertificate conditional. We add additional test cases ensuring that wildcard issuance is properly forbidden in all relevant scenarios, while allowing the existing test cases to validate that wildcard status doesn't affect non-wildcard certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add Wildcard allowance during signing operations When using sign-verbatim, sign-intermediate, or getting certificate generation parameters, set AllowWildcardCertificates to mirror existing policies. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-02-24 13:41:56 +00:00
})
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
Add role parameter to restrict issuance of wildcard certificates (#14238) * Add new AllowWildcardCertificate field to PKI role This field allows the PKI role to control whether or not issuance of wildcard certificates are allowed. We default (both on migration and new role creation) to the less secure true value for backwards compatibility with existing Vault versions. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sanitizedName to reducedName Per comment, this variable name was confusing during the reproduction and subsequent fix of the earlier vulnerability and associated bug report. Because the common name isn't necessarily _sanitized_ in any way (and indeed must be considered in relation to other parts or the whole), but portions of the entire name are removed, reducedName appears to make the most sense. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Enforce AllowWildcardCertificates during issuance This commit adds the bulk of correctly validating wildcard certificate Common Names during issuance according to RFC 6125 Section 6.4.3 semantics. As part of this, support for RFC 2818-conforming wildcard certificates (wherein there are almost no restrictions on issuance) has been removed. Note that this flag does take precedence over AllowAnyName, giving a little more safety in wildcard issuance in this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test cases to conform with RFC 6125 Test cases 19, 70+71, and 83+84 didn't conform with the RFC 6125, and so should've been rejected under strict conformance. For 70+71 and 83+84, we previously conditioned around the value of AllowSubdomains (allowing issuance when true), but they likely should've been rejected either way. Additionally, update the notes about globs matching wildcard certificates to notate this is indeed the case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Check AllowWildcardCertifciates in issuance tests This allows for regression tests to cover the new AllowWildcardCertificate conditional. We add additional test cases ensuring that wildcard issuance is properly forbidden in all relevant scenarios, while allowing the existing test cases to validate that wildcard status doesn't affect non-wildcard certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add Wildcard allowance during signing operations When using sign-verbatim, sign-intermediate, or getting certificate generation parameters, set AllowWildcardCertificates to mirror existing policies. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-02-24 13:41:56 +00:00
haveErr := err != nil || resp == nil
expectErr := !test.Issued
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
Add role parameter to restrict issuance of wildcard certificates (#14238) * Add new AllowWildcardCertificate field to PKI role This field allows the PKI role to control whether or not issuance of wildcard certificates are allowed. We default (both on migration and new role creation) to the less secure true value for backwards compatibility with existing Vault versions. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sanitizedName to reducedName Per comment, this variable name was confusing during the reproduction and subsequent fix of the earlier vulnerability and associated bug report. Because the common name isn't necessarily _sanitized_ in any way (and indeed must be considered in relation to other parts or the whole), but portions of the entire name are removed, reducedName appears to make the most sense. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Enforce AllowWildcardCertificates during issuance This commit adds the bulk of correctly validating wildcard certificate Common Names during issuance according to RFC 6125 Section 6.4.3 semantics. As part of this, support for RFC 2818-conforming wildcard certificates (wherein there are almost no restrictions on issuance) has been removed. Note that this flag does take precedence over AllowAnyName, giving a little more safety in wildcard issuance in this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test cases to conform with RFC 6125 Test cases 19, 70+71, and 83+84 didn't conform with the RFC 6125, and so should've been rejected under strict conformance. For 70+71 and 83+84, we previously conditioned around the value of AllowSubdomains (allowing issuance when true), but they likely should've been rejected either way. Additionally, update the notes about globs matching wildcard certificates to notate this is indeed the case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Check AllowWildcardCertifciates in issuance tests This allows for regression tests to cover the new AllowWildcardCertificate conditional. We add additional test cases ensuring that wildcard issuance is properly forbidden in all relevant scenarios, while allowing the existing test cases to validate that wildcard status doesn't affect non-wildcard certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add Wildcard allowance during signing operations When using sign-verbatim, sign-intermediate, or getting certificate generation parameters, set AllowWildcardCertificates to mirror existing policies. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-02-24 13:41:56 +00:00
if haveErr != expectErr {
t.Fatalf("issuance regression test [%d] failed: haveErr: %v, expectErr: %v, err: %v, resp: %v, test case: %v, role: %v", index, haveErr, expectErr, err, resp, test, role)
}
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
Add role parameter to restrict issuance of wildcard certificates (#14238) * Add new AllowWildcardCertificate field to PKI role This field allows the PKI role to control whether or not issuance of wildcard certificates are allowed. We default (both on migration and new role creation) to the less secure true value for backwards compatibility with existing Vault versions. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sanitizedName to reducedName Per comment, this variable name was confusing during the reproduction and subsequent fix of the earlier vulnerability and associated bug report. Because the common name isn't necessarily _sanitized_ in any way (and indeed must be considered in relation to other parts or the whole), but portions of the entire name are removed, reducedName appears to make the most sense. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Enforce AllowWildcardCertificates during issuance This commit adds the bulk of correctly validating wildcard certificate Common Names during issuance according to RFC 6125 Section 6.4.3 semantics. As part of this, support for RFC 2818-conforming wildcard certificates (wherein there are almost no restrictions on issuance) has been removed. Note that this flag does take precedence over AllowAnyName, giving a little more safety in wildcard issuance in this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test cases to conform with RFC 6125 Test cases 19, 70+71, and 83+84 didn't conform with the RFC 6125, and so should've been rejected under strict conformance. For 70+71 and 83+84, we previously conditioned around the value of AllowSubdomains (allowing issuance when true), but they likely should've been rejected either way. Additionally, update the notes about globs matching wildcard certificates to notate this is indeed the case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Check AllowWildcardCertifciates in issuance tests This allows for regression tests to cover the new AllowWildcardCertificate conditional. We add additional test cases ensuring that wildcard issuance is properly forbidden in all relevant scenarios, while allowing the existing test cases to validate that wildcard status doesn't affect non-wildcard certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add Wildcard allowance during signing operations When using sign-verbatim, sign-intermediate, or getting certificate generation parameters, set AllowWildcardCertificates to mirror existing policies. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-02-24 13:41:56 +00:00
tested += 1
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
}
}
}
}
}
return tested
}
func TestBackend_Roles_IssuanceRegression(t *testing.T) {
t.Parallel()
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
// Regression testing of role's issuance policy.
testCases := []IssuanceRegression{
Add role parameter to restrict issuance of wildcard certificates (#14238) * Add new AllowWildcardCertificate field to PKI role This field allows the PKI role to control whether or not issuance of wildcard certificates are allowed. We default (both on migration and new role creation) to the less secure true value for backwards compatibility with existing Vault versions. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sanitizedName to reducedName Per comment, this variable name was confusing during the reproduction and subsequent fix of the earlier vulnerability and associated bug report. Because the common name isn't necessarily _sanitized_ in any way (and indeed must be considered in relation to other parts or the whole), but portions of the entire name are removed, reducedName appears to make the most sense. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Enforce AllowWildcardCertificates during issuance This commit adds the bulk of correctly validating wildcard certificate Common Names during issuance according to RFC 6125 Section 6.4.3 semantics. As part of this, support for RFC 2818-conforming wildcard certificates (wherein there are almost no restrictions on issuance) has been removed. Note that this flag does take precedence over AllowAnyName, giving a little more safety in wildcard issuance in this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test cases to conform with RFC 6125 Test cases 19, 70+71, and 83+84 didn't conform with the RFC 6125, and so should've been rejected under strict conformance. For 70+71 and 83+84, we previously conditioned around the value of AllowSubdomains (allowing issuance when true), but they likely should've been rejected either way. Additionally, update the notes about globs matching wildcard certificates to notate this is indeed the case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Check AllowWildcardCertifciates in issuance tests This allows for regression tests to cover the new AllowWildcardCertificate conditional. We add additional test cases ensuring that wildcard issuance is properly forbidden in all relevant scenarios, while allowing the existing test cases to validate that wildcard status doesn't affect non-wildcard certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add Wildcard allowance during signing operations When using sign-verbatim, sign-intermediate, or getting certificate generation parameters, set AllowWildcardCertificates to mirror existing policies. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-02-24 13:41:56 +00:00
// allowed, bare, glob, subdomains, localhost, wildcards, cn, issued
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
Add role parameter to restrict issuance of wildcard certificates (#14238) * Add new AllowWildcardCertificate field to PKI role This field allows the PKI role to control whether or not issuance of wildcard certificates are allowed. We default (both on migration and new role creation) to the less secure true value for backwards compatibility with existing Vault versions. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sanitizedName to reducedName Per comment, this variable name was confusing during the reproduction and subsequent fix of the earlier vulnerability and associated bug report. Because the common name isn't necessarily _sanitized_ in any way (and indeed must be considered in relation to other parts or the whole), but portions of the entire name are removed, reducedName appears to make the most sense. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Enforce AllowWildcardCertificates during issuance This commit adds the bulk of correctly validating wildcard certificate Common Names during issuance according to RFC 6125 Section 6.4.3 semantics. As part of this, support for RFC 2818-conforming wildcard certificates (wherein there are almost no restrictions on issuance) has been removed. Note that this flag does take precedence over AllowAnyName, giving a little more safety in wildcard issuance in this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test cases to conform with RFC 6125 Test cases 19, 70+71, and 83+84 didn't conform with the RFC 6125, and so should've been rejected under strict conformance. For 70+71 and 83+84, we previously conditioned around the value of AllowSubdomains (allowing issuance when true), but they likely should've been rejected either way. Additionally, update the notes about globs matching wildcard certificates to notate this is indeed the case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Check AllowWildcardCertifciates in issuance tests This allows for regression tests to cover the new AllowWildcardCertificate conditional. We add additional test cases ensuring that wildcard issuance is properly forbidden in all relevant scenarios, while allowing the existing test cases to validate that wildcard status doesn't affect non-wildcard certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add Wildcard allowance during signing operations When using sign-verbatim, sign-intermediate, or getting certificate generation parameters, set AllowWildcardCertificates to mirror existing policies. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-02-24 13:41:56 +00:00
// === Globs not allowed but used === //
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
// Allowed contains globs, but globbing not allowed, resulting in all
// issuances failing. Note that tests against issuing a wildcard with
// a bare domain will be covered later.
/* 0 */ {[]string{"*.*.foo"}, MAny, MFalse, MAny, MAny, MAny, nil, "baz.fud.bar.foo", false},
/* 1 */ {[]string{"*.*.foo"}, MAny, MFalse, MAny, MAny, MAny, nil, "*.fud.bar.foo", false},
/* 2 */ {[]string{"*.*.foo"}, MAny, MFalse, MAny, MAny, MAny, nil, "fud.bar.foo", false},
/* 3 */ {[]string{"*.*.foo"}, MAny, MFalse, MAny, MAny, MAny, nil, "*.bar.foo", false},
/* 4 */ {[]string{"*.*.foo"}, MAny, MFalse, MAny, MAny, MAny, nil, "bar.foo", false},
/* 5 */ {[]string{"*.*.foo"}, MAny, MFalse, MAny, MAny, MAny, nil, "*.foo", false},
/* 6 */ {[]string{"*.foo"}, MAny, MFalse, MAny, MAny, MAny, nil, "foo", false},
/* 7 */ {[]string{"*.foo"}, MAny, MFalse, MAny, MAny, MAny, nil, "baz.fud.bar.foo", false},
/* 8 */ {[]string{"*.foo"}, MAny, MFalse, MAny, MAny, MAny, nil, "*.fud.bar.foo", false},
/* 9 */ {[]string{"*.foo"}, MAny, MFalse, MAny, MAny, MAny, nil, "fud.bar.foo", false},
/* 10 */ {[]string{"*.foo"}, MAny, MFalse, MAny, MAny, MAny, nil, "*.bar.foo", false},
/* 11 */ {[]string{"*.foo"}, MAny, MFalse, MAny, MAny, MAny, nil, "bar.foo", false},
/* 12 */ {[]string{"*.foo"}, MAny, MFalse, MAny, MAny, MAny, nil, "foo", false},
Add role parameter to restrict issuance of wildcard certificates (#14238) * Add new AllowWildcardCertificate field to PKI role This field allows the PKI role to control whether or not issuance of wildcard certificates are allowed. We default (both on migration and new role creation) to the less secure true value for backwards compatibility with existing Vault versions. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sanitizedName to reducedName Per comment, this variable name was confusing during the reproduction and subsequent fix of the earlier vulnerability and associated bug report. Because the common name isn't necessarily _sanitized_ in any way (and indeed must be considered in relation to other parts or the whole), but portions of the entire name are removed, reducedName appears to make the most sense. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Enforce AllowWildcardCertificates during issuance This commit adds the bulk of correctly validating wildcard certificate Common Names during issuance according to RFC 6125 Section 6.4.3 semantics. As part of this, support for RFC 2818-conforming wildcard certificates (wherein there are almost no restrictions on issuance) has been removed. Note that this flag does take precedence over AllowAnyName, giving a little more safety in wildcard issuance in this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test cases to conform with RFC 6125 Test cases 19, 70+71, and 83+84 didn't conform with the RFC 6125, and so should've been rejected under strict conformance. For 70+71 and 83+84, we previously conditioned around the value of AllowSubdomains (allowing issuance when true), but they likely should've been rejected either way. Additionally, update the notes about globs matching wildcard certificates to notate this is indeed the case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Check AllowWildcardCertifciates in issuance tests This allows for regression tests to cover the new AllowWildcardCertificate conditional. We add additional test cases ensuring that wildcard issuance is properly forbidden in all relevant scenarios, while allowing the existing test cases to validate that wildcard status doesn't affect non-wildcard certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add Wildcard allowance during signing operations When using sign-verbatim, sign-intermediate, or getting certificate generation parameters, set AllowWildcardCertificates to mirror existing policies. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-02-24 13:41:56 +00:00
// === Localhost sanity === //
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
// Localhost forbidden, not matching allowed domains -> not issued
/* 13 */ {[]string{"*.*.foo"}, MAny, MAny, MAny, MFalse, MAny, nil, "localhost", false},
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
// Localhost allowed, not matching allowed domains -> issued
/* 14 */ {[]string{"*.*.foo"}, MAny, MAny, MAny, MTrue, MAny, nil, "localhost", true},
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
// Localhost allowed via allowed domains (and bare allowed), not by AllowLocalhost -> issued
/* 15 */ {[]string{"localhost"}, MTrue, MAny, MAny, MFalse, MAny, nil, "localhost", true},
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
// Localhost allowed via allowed domains (and bare not allowed), not by AllowLocalhost -> not issued
/* 16 */ {[]string{"localhost"}, MFalse, MAny, MAny, MFalse, MAny, nil, "localhost", false},
Add role parameter to restrict issuance of wildcard certificates (#14238) * Add new AllowWildcardCertificate field to PKI role This field allows the PKI role to control whether or not issuance of wildcard certificates are allowed. We default (both on migration and new role creation) to the less secure true value for backwards compatibility with existing Vault versions. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sanitizedName to reducedName Per comment, this variable name was confusing during the reproduction and subsequent fix of the earlier vulnerability and associated bug report. Because the common name isn't necessarily _sanitized_ in any way (and indeed must be considered in relation to other parts or the whole), but portions of the entire name are removed, reducedName appears to make the most sense. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Enforce AllowWildcardCertificates during issuance This commit adds the bulk of correctly validating wildcard certificate Common Names during issuance according to RFC 6125 Section 6.4.3 semantics. As part of this, support for RFC 2818-conforming wildcard certificates (wherein there are almost no restrictions on issuance) has been removed. Note that this flag does take precedence over AllowAnyName, giving a little more safety in wildcard issuance in this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test cases to conform with RFC 6125 Test cases 19, 70+71, and 83+84 didn't conform with the RFC 6125, and so should've been rejected under strict conformance. For 70+71 and 83+84, we previously conditioned around the value of AllowSubdomains (allowing issuance when true), but they likely should've been rejected either way. Additionally, update the notes about globs matching wildcard certificates to notate this is indeed the case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Check AllowWildcardCertifciates in issuance tests This allows for regression tests to cover the new AllowWildcardCertificate conditional. We add additional test cases ensuring that wildcard issuance is properly forbidden in all relevant scenarios, while allowing the existing test cases to validate that wildcard status doesn't affect non-wildcard certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add Wildcard allowance during signing operations When using sign-verbatim, sign-intermediate, or getting certificate generation parameters, set AllowWildcardCertificates to mirror existing policies. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-02-24 13:41:56 +00:00
// Localhost allowed via allowed domains (but bare not allowed), and by AllowLocalhost -> issued
/* 17 */ {[]string{"localhost"}, MFalse, MAny, MAny, MTrue, MAny, nil, "localhost", true},
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
Add role parameter to restrict issuance of wildcard certificates (#14238) * Add new AllowWildcardCertificate field to PKI role This field allows the PKI role to control whether or not issuance of wildcard certificates are allowed. We default (both on migration and new role creation) to the less secure true value for backwards compatibility with existing Vault versions. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sanitizedName to reducedName Per comment, this variable name was confusing during the reproduction and subsequent fix of the earlier vulnerability and associated bug report. Because the common name isn't necessarily _sanitized_ in any way (and indeed must be considered in relation to other parts or the whole), but portions of the entire name are removed, reducedName appears to make the most sense. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Enforce AllowWildcardCertificates during issuance This commit adds the bulk of correctly validating wildcard certificate Common Names during issuance according to RFC 6125 Section 6.4.3 semantics. As part of this, support for RFC 2818-conforming wildcard certificates (wherein there are almost no restrictions on issuance) has been removed. Note that this flag does take precedence over AllowAnyName, giving a little more safety in wildcard issuance in this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test cases to conform with RFC 6125 Test cases 19, 70+71, and 83+84 didn't conform with the RFC 6125, and so should've been rejected under strict conformance. For 70+71 and 83+84, we previously conditioned around the value of AllowSubdomains (allowing issuance when true), but they likely should've been rejected either way. Additionally, update the notes about globs matching wildcard certificates to notate this is indeed the case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Check AllowWildcardCertifciates in issuance tests This allows for regression tests to cover the new AllowWildcardCertificate conditional. We add additional test cases ensuring that wildcard issuance is properly forbidden in all relevant scenarios, while allowing the existing test cases to validate that wildcard status doesn't affect non-wildcard certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add Wildcard allowance during signing operations When using sign-verbatim, sign-intermediate, or getting certificate generation parameters, set AllowWildcardCertificates to mirror existing policies. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-02-24 13:41:56 +00:00
// === Bare wildcard issuance == //
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
// allowed_domains contains one or more wildcards and bare domains allowed,
// resulting in the cert being issued.
/* 18 */ {[]string{"*.foo"}, MTrue, MAny, MAny, MAny, MTrue, nil, "*.foo", true},
/* 19 */ {[]string{"*.*.foo"}, MTrue, MAny, MAny, MAny, MAny, nil, "*.*.foo", false}, // Does not conform to RFC 6125
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
// === Double Leading Glob Testing === //
// Allowed contains globs, but glob allowed so certain matches work.
// The value of bare and localhost does not impact these results.
/* 20 */ {[]string{"*.*.foo"}, MAny, MTrue, MFalse, MAny, MAny, nil, "baz.fud.bar.foo", true}, // glob domains allow infinite subdomains
/* 21 */ {[]string{"*.*.foo"}, MAny, MTrue, MFalse, MAny, MTrue, nil, "*.fud.bar.foo", true}, // glob domain allows wildcard of subdomains
/* 22 */ {[]string{"*.*.foo"}, MAny, MTrue, MFalse, MAny, MAny, nil, "fud.bar.foo", true},
/* 23 */ {[]string{"*.*.foo"}, MAny, MTrue, MFalse, MAny, MTrue, nil, "*.bar.foo", true}, // Regression fix: Vault#13530
/* 24 */ {[]string{"*.*.foo"}, MAny, MTrue, MFalse, MAny, MAny, nil, "bar.foo", false},
/* 25 */ {[]string{"*.*.foo"}, MAny, MTrue, MFalse, MAny, MAny, nil, "*.foo", false},
/* 26 */ {[]string{"*.*.foo"}, MAny, MTrue, MFalse, MAny, MAny, nil, "foo", false},
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
// Allowed contains globs, but glob and subdomain both work, so we expect
// wildcard issuance to work as well. The value of bare and localhost does
// not impact these results.
/* 27 */ {[]string{"*.*.foo"}, MAny, MTrue, MTrue, MAny, MAny, nil, "baz.fud.bar.foo", true},
/* 28 */ {[]string{"*.*.foo"}, MAny, MTrue, MTrue, MAny, MTrue, nil, "*.fud.bar.foo", true},
/* 29 */ {[]string{"*.*.foo"}, MAny, MTrue, MTrue, MAny, MAny, nil, "fud.bar.foo", true},
/* 30 */ {[]string{"*.*.foo"}, MAny, MTrue, MTrue, MAny, MTrue, nil, "*.bar.foo", true}, // Regression fix: Vault#13530
/* 31 */ {[]string{"*.*.foo"}, MAny, MTrue, MTrue, MAny, MAny, nil, "bar.foo", false},
/* 32 */ {[]string{"*.*.foo"}, MAny, MTrue, MTrue, MAny, MAny, nil, "*.foo", false},
/* 33 */ {[]string{"*.*.foo"}, MAny, MTrue, MTrue, MAny, MAny, nil, "foo", false},
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
// === Single Leading Glob Testing === //
// Allowed contains globs, but glob allowed so certain matches work.
// The value of bare and localhost does not impact these results.
/* 34 */ {[]string{"*.foo"}, MAny, MTrue, MFalse, MAny, MAny, nil, "baz.fud.bar.foo", true}, // glob domains allow infinite subdomains
/* 35 */ {[]string{"*.foo"}, MAny, MTrue, MFalse, MAny, MTrue, nil, "*.fud.bar.foo", true}, // glob domain allows wildcard of subdomains
/* 36 */ {[]string{"*.foo"}, MAny, MTrue, MFalse, MAny, MAny, nil, "fud.bar.foo", true}, // glob domains allow infinite subdomains
/* 37 */ {[]string{"*.foo"}, MAny, MTrue, MFalse, MAny, MTrue, nil, "*.bar.foo", true}, // glob domain allows wildcards of subdomains
/* 38 */ {[]string{"*.foo"}, MAny, MTrue, MFalse, MAny, MAny, nil, "bar.foo", true},
/* 39 */ {[]string{"*.foo"}, MAny, MTrue, MFalse, MAny, MAny, nil, "foo", false},
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
// Allowed contains globs, but glob and subdomain both work, so we expect
// wildcard issuance to work as well. The value of bare and localhost does
// not impact these results.
/* 40 */ {[]string{"*.foo"}, MAny, MTrue, MTrue, MAny, MAny, nil, "baz.fud.bar.foo", true},
/* 41 */ {[]string{"*.foo"}, MAny, MTrue, MTrue, MAny, MTrue, nil, "*.fud.bar.foo", true},
/* 42 */ {[]string{"*.foo"}, MAny, MTrue, MTrue, MAny, MAny, nil, "fud.bar.foo", true},
/* 43 */ {[]string{"*.foo"}, MAny, MTrue, MTrue, MAny, MTrue, nil, "*.bar.foo", true},
/* 44 */ {[]string{"*.foo"}, MAny, MTrue, MTrue, MAny, MAny, nil, "bar.foo", true},
/* 45 */ {[]string{"*.foo"}, MAny, MTrue, MTrue, MAny, MAny, nil, "foo", false},
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
// === Only base domain name === //
// Allowed contains only domain components, but subdomains not allowed. This
// results in most issuances failing unless we allow bare domains, in which
// case only the final issuance for "foo" will succeed.
/* 46 */ {[]string{"foo"}, MAny, MAny, MFalse, MAny, MAny, nil, "baz.fud.bar.foo", false},
/* 47 */ {[]string{"foo"}, MAny, MAny, MFalse, MAny, MAny, nil, "*.fud.bar.foo", false},
/* 48 */ {[]string{"foo"}, MAny, MAny, MFalse, MAny, MAny, nil, "fud.bar.foo", false},
/* 49 */ {[]string{"foo"}, MAny, MAny, MFalse, MAny, MAny, nil, "*.bar.foo", false},
/* 50 */ {[]string{"foo"}, MAny, MAny, MFalse, MAny, MAny, nil, "bar.foo", false},
/* 51 */ {[]string{"foo"}, MAny, MAny, MFalse, MAny, MAny, nil, "*.foo", false},
/* 52 */ {[]string{"foo"}, MFalse, MAny, MFalse, MAny, MAny, nil, "foo", false},
/* 53 */ {[]string{"foo"}, MTrue, MAny, MFalse, MAny, MAny, nil, "foo", true},
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
// Allowed contains only domain components, and subdomains are now allowed.
// This results in most issuances succeeding, with the exception of the
// base foo, which is still governed by base's value.
/* 54 */ {[]string{"foo"}, MAny, MAny, MTrue, MAny, MAny, nil, "baz.fud.bar.foo", true},
/* 55 */ {[]string{"foo"}, MAny, MAny, MTrue, MAny, MTrue, nil, "*.fud.bar.foo", true},
/* 56 */ {[]string{"foo"}, MAny, MAny, MTrue, MAny, MAny, nil, "fud.bar.foo", true},
/* 57 */ {[]string{"foo"}, MAny, MAny, MTrue, MAny, MTrue, nil, "*.bar.foo", true},
/* 58 */ {[]string{"foo"}, MAny, MAny, MTrue, MAny, MAny, nil, "bar.foo", true},
/* 59 */ {[]string{"foo"}, MAny, MAny, MTrue, MAny, MTrue, nil, "*.foo", true},
/* 60 */ {[]string{"foo"}, MAny, MAny, MTrue, MAny, MTrue, nil, "x*x.foo", true}, // internal wildcards should be allowed per RFC 6125/6.4.3
/* 61 */ {[]string{"foo"}, MAny, MAny, MTrue, MAny, MTrue, nil, "*x.foo", true}, // prefix wildcards should be allowed per RFC 6125/6.4.3
/* 62 */ {[]string{"foo"}, MAny, MAny, MTrue, MAny, MTrue, nil, "x*.foo", true}, // suffix wildcards should be allowed per RFC 6125/6.4.3
/* 63 */ {[]string{"foo"}, MFalse, MAny, MTrue, MAny, MAny, nil, "foo", false},
/* 64 */ {[]string{"foo"}, MTrue, MAny, MTrue, MAny, MAny, nil, "foo", true},
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
// === Internal Glob Matching === //
// Basic glob matching requirements
/* 65 */ {[]string{"x*x.foo"}, MAny, MTrue, MAny, MAny, MAny, nil, "xerox.foo", true},
/* 66 */ {[]string{"x*x.foo"}, MAny, MTrue, MAny, MAny, MAny, nil, "xylophone.files.pyrex.foo", true}, // globs can match across subdomains
/* 67 */ {[]string{"x*x.foo"}, MAny, MTrue, MAny, MAny, MAny, nil, "xercex.bar.foo", false}, // x.foo isn't matched
/* 68 */ {[]string{"x*x.foo"}, MAny, MTrue, MAny, MAny, MAny, nil, "bar.foo", false}, // x*x isn't matched.
/* 69 */ {[]string{"x*x.foo"}, MAny, MTrue, MAny, MAny, MAny, nil, "*.foo", false}, // unrelated wildcard
/* 70 */ {[]string{"x*x.foo"}, MAny, MTrue, MAny, MAny, MAny, nil, "*.x*x.foo", false}, // Does not conform to RFC 6125
/* 71 */ {[]string{"x*x.foo"}, MAny, MTrue, MAny, MAny, MAny, nil, "*.xyx.foo", false}, // Globs and Subdomains do not layer per docs.
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
// Various requirements around x*x.foo wildcard matching.
/* 72 */ {[]string{"x*x.foo"}, MFalse, MFalse, MAny, MAny, MAny, nil, "x*x.foo", false}, // base disabled, shouldn't match wildcard
/* 73 */ {[]string{"x*x.foo"}, MFalse, MTrue, MAny, MAny, MTrue, nil, "x*x.foo", true}, // base disallowed, but globbing allowed and should match
/* 74 */ {[]string{"x*x.foo"}, MTrue, MAny, MAny, MAny, MTrue, nil, "x*x.foo", true}, // base allowed, should match wildcard
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
// Basic glob matching requirements with internal dots.
/* 75 */ {[]string{"x.*.x.foo"}, MAny, MTrue, MAny, MAny, MAny, nil, "xerox.foo", false}, // missing dots
/* 76 */ {[]string{"x.*.x.foo"}, MAny, MTrue, MAny, MAny, MAny, nil, "x.ero.x.foo", true},
/* 77 */ {[]string{"x.*.x.foo"}, MAny, MTrue, MAny, MAny, MAny, nil, "xylophone.files.pyrex.foo", false}, // missing dots
/* 78 */ {[]string{"x.*.x.foo"}, MAny, MTrue, MAny, MAny, MAny, nil, "x.ylophone.files.pyre.x.foo", true}, // globs can match across subdomains
/* 79 */ {[]string{"x.*.x.foo"}, MAny, MTrue, MAny, MAny, MAny, nil, "xercex.bar.foo", false}, // x.foo isn't matched
/* 80 */ {[]string{"x.*.x.foo"}, MAny, MTrue, MAny, MAny, MAny, nil, "bar.foo", false}, // x.*.x isn't matched.
/* 81 */ {[]string{"x.*.x.foo"}, MAny, MTrue, MAny, MAny, MAny, nil, "*.foo", false}, // unrelated wildcard
/* 82 */ {[]string{"x.*.x.foo"}, MAny, MTrue, MAny, MAny, MAny, nil, "*.x.*.x.foo", false}, // Does not conform to RFC 6125
/* 83 */ {[]string{"x.*.x.foo"}, MAny, MTrue, MAny, MAny, MAny, nil, "*.x.y.x.foo", false}, // Globs and Subdomains do not layer per docs.
Add role parameter to restrict issuance of wildcard certificates (#14238) * Add new AllowWildcardCertificate field to PKI role This field allows the PKI role to control whether or not issuance of wildcard certificates are allowed. We default (both on migration and new role creation) to the less secure true value for backwards compatibility with existing Vault versions. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sanitizedName to reducedName Per comment, this variable name was confusing during the reproduction and subsequent fix of the earlier vulnerability and associated bug report. Because the common name isn't necessarily _sanitized_ in any way (and indeed must be considered in relation to other parts or the whole), but portions of the entire name are removed, reducedName appears to make the most sense. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Enforce AllowWildcardCertificates during issuance This commit adds the bulk of correctly validating wildcard certificate Common Names during issuance according to RFC 6125 Section 6.4.3 semantics. As part of this, support for RFC 2818-conforming wildcard certificates (wherein there are almost no restrictions on issuance) has been removed. Note that this flag does take precedence over AllowAnyName, giving a little more safety in wildcard issuance in this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test cases to conform with RFC 6125 Test cases 19, 70+71, and 83+84 didn't conform with the RFC 6125, and so should've been rejected under strict conformance. For 70+71 and 83+84, we previously conditioned around the value of AllowSubdomains (allowing issuance when true), but they likely should've been rejected either way. Additionally, update the notes about globs matching wildcard certificates to notate this is indeed the case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Check AllowWildcardCertifciates in issuance tests This allows for regression tests to cover the new AllowWildcardCertificate conditional. We add additional test cases ensuring that wildcard issuance is properly forbidden in all relevant scenarios, while allowing the existing test cases to validate that wildcard status doesn't affect non-wildcard certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add Wildcard allowance during signing operations When using sign-verbatim, sign-intermediate, or getting certificate generation parameters, set AllowWildcardCertificates to mirror existing policies. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-02-24 13:41:56 +00:00
// === Wildcard restriction testing === //
/* 84 */ {[]string{"*.foo"}, MAny, MTrue, MFalse, MAny, MFalse, nil, "*.fud.bar.foo", false}, // glob domain allows wildcard of subdomains
/* 85 */ {[]string{"*.foo"}, MAny, MTrue, MFalse, MAny, MFalse, nil, "*.bar.foo", false}, // glob domain allows wildcards of subdomains
/* 86 */ {[]string{"foo"}, MAny, MAny, MTrue, MAny, MFalse, nil, "*.fud.bar.foo", false},
/* 87 */ {[]string{"foo"}, MAny, MAny, MTrue, MAny, MFalse, nil, "*.bar.foo", false},
/* 88 */ {[]string{"foo"}, MAny, MAny, MTrue, MAny, MFalse, nil, "*.foo", false},
/* 89 */ {[]string{"foo"}, MAny, MAny, MTrue, MAny, MFalse, nil, "x*x.foo", false},
/* 90 */ {[]string{"foo"}, MAny, MAny, MTrue, MAny, MFalse, nil, "*x.foo", false},
/* 91 */ {[]string{"foo"}, MAny, MAny, MTrue, MAny, MFalse, nil, "x*.foo", false},
/* 92 */ {[]string{"x*x.foo"}, MTrue, MAny, MAny, MAny, MFalse, nil, "x*x.foo", false},
/* 93 */ {[]string{"*.foo"}, MFalse, MFalse, MAny, MAny, MAny, nil, "*.foo", false}, // Bare and globs forbidden despite (potentially) allowing wildcards.
/* 94 */ {[]string{"x.*.x.foo"}, MAny, MAny, MAny, MAny, MAny, nil, "x.*.x.foo", false}, // Does not conform to RFC 6125
// === CN validation allowances === //
/* 95 */ {[]string{"foo"}, MAny, MAny, MAny, MAny, MAny, []string{"disabled"}, "*.fud.bar.foo", true},
/* 96 */ {[]string{"foo"}, MAny, MAny, MAny, MAny, MAny, []string{"disabled"}, "*.fud.*.foo", true},
/* 97 */ {[]string{"foo"}, MAny, MAny, MAny, MAny, MAny, []string{"disabled"}, "*.bar.*.bar", true},
/* 98 */ {[]string{"foo"}, MAny, MAny, MAny, MAny, MAny, []string{"disabled"}, "foo@foo", true},
/* 99 */ {[]string{"foo"}, MAny, MAny, MAny, MAny, MAny, []string{"disabled"}, "foo@foo@foo", true},
/* 100 */ {[]string{"foo"}, MAny, MAny, MAny, MAny, MAny, []string{"disabled"}, "bar@bar@bar", true},
/* 101 */ {[]string{"foo"}, MTrue, MTrue, MTrue, MTrue, MTrue, []string{"email"}, "bar@bar@bar", false},
/* 102 */ {[]string{"foo"}, MTrue, MTrue, MTrue, MTrue, MTrue, []string{"email"}, "bar@bar", false},
/* 103 */ {[]string{"foo"}, MTrue, MTrue, MTrue, MTrue, MTrue, []string{"email"}, "bar@foo", true},
/* 104 */ {[]string{"foo"}, MTrue, MTrue, MTrue, MTrue, MTrue, []string{"hostname"}, "bar@foo", false},
/* 105 */ {[]string{"foo"}, MTrue, MTrue, MTrue, MTrue, MTrue, []string{"hostname"}, "bar@bar", false},
/* 106 */ {[]string{"foo"}, MTrue, MTrue, MTrue, MTrue, MTrue, []string{"hostname"}, "bar.foo", true},
/* 107 */ {[]string{"foo"}, MTrue, MTrue, MTrue, MTrue, MTrue, []string{"hostname"}, "bar.bar", false},
/* 108 */ {[]string{"foo"}, MTrue, MTrue, MTrue, MTrue, MTrue, []string{"email"}, "bar.foo", false},
/* 109 */ {[]string{"foo"}, MTrue, MTrue, MTrue, MTrue, MTrue, []string{"email"}, "bar.bar", false},
}
if len(testCases) != 110 {
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
t.Fatalf("misnumbered test case entries will make it hard to find bugs: %v", len(testCases))
}
b, s := CreateBackendWithStorage(t)
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
// We need a RSA key so all signature sizes are valid with it.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err := CBWrite(b, s, "root/generate/exported", map[string]interface{}{
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
"common_name": "myvault.com",
"ttl": "128h",
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
"key_type": "rsa",
"key_bits": 2048,
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
})
if err != nil {
t.Fatal(err)
}
if resp == nil {
t.Fatal("expected ca info")
}
tested := 0
for index, test := range testCases {
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
tested += RoleIssuanceRegressionHelper(t, b, s, index, test)
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
}
t.Logf("Issuance regression expanded matrix test scenarios: %d", tested)
Fix broken interactions between glob_domains and wildcards (#14235) * Allow issuance of wildcard via glob match From Vault v1.8.0 onwards, we would incorrectly disallow issuance of a wildcard certificate when allow_glob_domain was enabled with a multi-part glob domain in allowed_domains (such as *.*.foo) when attempting to issue a wildcard for a subdomain (such as *.bar.foo). This fixes that by reverting an errant change in the case insensitivity patch. Here, when validating against a very powerful glob construct, we leave the wildcard prefix (*.) from the raw common_name element, to allow multi-part globs to match wildcard entries. It is important to note that "sanitizedName" is an incorrect variable naming here. Wildcard parsing (per RFC 6125 which supercedes RFC 2818) must be in the left-most segment of the domain, but we lack validation to ensure no internal wildcards exist. Additionally per item 3 of section 6.4.3 of RFC 6125, wildcards MAY be internal to a domain segment, in which case sanitizedName again leaves the wildcard in place. Resolves: #13530 Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove duplicate email address check As pointed out by Steven Clark (author of the removed conditional in 70012cd865b3dcdab376dba0c0e0abc88c48f508), this is duplicate from the now-reintroduced comparison against name (versus the erroneous sanitizedName at the time of his commit). This is a reversion of the changes to builtin/logical/pki/cert_util.go, but keeping the additional valuable test cases. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add multi-dimensional PKI role issuance tests This commit introduces multi-dimensional testing of PKI secrets engine's role-based certificate issuance with the intent of preventing future regressions. Here, dimensions of testing include: - AllowedDomains to decide which domains are approved for issuance, - AllowBareDomains to decide if raw entries of AllowedDomains are permitted, - AllowGlobDomains to decide if glob patterns in AllowedDomains are parsed, - AllowSubdomains to decide if subdomains of AllowedDomains are permitted, - AllowLocalhost to decide if localhost identifiers are permitted, and - CommonName of the certificate to request. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-02-23 21:44:09 +00:00
}
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
type KeySizeRegression struct {
// Values reused for both Role and CA configuration.
RoleKeyType string
RoleKeyBits []int
// Signature Bits presently is only specified on the role.
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
RoleSignatureBits []int
Add PSS support to PKI Secrets Engine (#16519) * Add PSS signature support to Vault PKI engine Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use issuer's RevocationSigAlg for CRL signing We introduce a new parameter on issuers, revocation_signature_algorithm to control the signature algorithm used during CRL signing. This is because the SignatureAlgorithm value from the certificate itself is incorrect for this purpose: a RSA root could sign an ECDSA intermediate with say, SHA256WithRSA, but when the intermediate goes to sign a CRL, it must use ECDSAWithSHA256 or equivalent instead of SHA256WithRSA. When coupled with support for PSS-only keys, allowing the user to set the signature algorithm value as desired seems like the best approach. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add use_pss, revocation_signature_algorithm docs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add PSS to signature role issuance test matrix Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow roots to self-identify revocation alg When using PSS support with a managed key, sometimes the underlying device will not support PKCS#1v1.5 signatures. This results in CRL building failing, unless we update the entry's signature algorithm prior to building the CRL for the new root. With a RSA-type key and use_pss=true, we use the signature bits value to decide which hash function to use for PSS support. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add clearer error message on failed import When CRL building fails during cert/key import, due to PSS failures, give a better indication to the user that import succeeded its just CRL building that failed. This tells them the parameter to adjust on the issuer and warns that CRL building will fail until this is fixed. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add case insensitive SigAlgo matching Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert UsePSS back to regular bool Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor PSS->certTemplate into helper function Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Proper string output on rev_sig_alg display Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Copy root's SignatureAlgorithm for CRL building Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-03 16:42:24 +00:00
RoleUsePSS bool
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
// These are tuples; must be of the same length.
TestKeyTypes []string
TestKeyBits []int
// All of the above key types/sizes must pass or fail together.
ExpectError bool
}
func (k KeySizeRegression) KeyTypeValues() []string {
if k.RoleKeyType == "any" {
return []string{"rsa", "ec", "ed25519"}
}
return []string{k.RoleKeyType}
}
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
func RoleKeySizeRegressionHelper(t *testing.T, b *backend, s logical.Storage, index int, test KeySizeRegression) int {
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
tested := 0
for _, caKeyType := range test.KeyTypeValues() {
for _, caKeyBits := range test.RoleKeyBits {
// Generate a new CA key.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err := CBWrite(b, s, "root/generate/exported", map[string]interface{}{
"common_name": "myvault.com",
"ttl": "128h",
"key_type": caKeyType,
"key_bits": caKeyBits,
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
})
if err != nil {
t.Fatal(err)
}
if resp == nil {
t.Fatal("expected ca info")
}
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
for _, roleKeyBits := range test.RoleKeyBits {
for _, roleSignatureBits := range test.RoleSignatureBits {
role := fmt.Sprintf("key-size-regression-%d-keytype-%v-keybits-%d-signature-bits-%d", index, test.RoleKeyType, roleKeyBits, roleSignatureBits)
_, err := CBWrite(b, s, "roles/"+role, map[string]interface{}{
"key_type": test.RoleKeyType,
"key_bits": roleKeyBits,
"signature_bits": roleSignatureBits,
Add PSS support to PKI Secrets Engine (#16519) * Add PSS signature support to Vault PKI engine Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use issuer's RevocationSigAlg for CRL signing We introduce a new parameter on issuers, revocation_signature_algorithm to control the signature algorithm used during CRL signing. This is because the SignatureAlgorithm value from the certificate itself is incorrect for this purpose: a RSA root could sign an ECDSA intermediate with say, SHA256WithRSA, but when the intermediate goes to sign a CRL, it must use ECDSAWithSHA256 or equivalent instead of SHA256WithRSA. When coupled with support for PSS-only keys, allowing the user to set the signature algorithm value as desired seems like the best approach. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add use_pss, revocation_signature_algorithm docs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add PSS to signature role issuance test matrix Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow roots to self-identify revocation alg When using PSS support with a managed key, sometimes the underlying device will not support PKCS#1v1.5 signatures. This results in CRL building failing, unless we update the entry's signature algorithm prior to building the CRL for the new root. With a RSA-type key and use_pss=true, we use the signature bits value to decide which hash function to use for PSS support. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add clearer error message on failed import When CRL building fails during cert/key import, due to PSS failures, give a better indication to the user that import succeeded its just CRL building that failed. This tells them the parameter to adjust on the issuer and warns that CRL building will fail until this is fixed. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add case insensitive SigAlgo matching Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert UsePSS back to regular bool Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor PSS->certTemplate into helper function Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Proper string output on rev_sig_alg display Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Copy root's SignatureAlgorithm for CRL building Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-03 16:42:24 +00:00
"use_pss": test.RoleUsePSS,
})
if err != nil {
t.Fatal(err)
}
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
for index, keyType := range test.TestKeyTypes {
keyBits := test.TestKeyBits[index]
_, _, csrPem := generateCSR(t, &x509.CertificateRequest{
Subject: pkix.Name{
CommonName: "localhost",
},
}, keyType, keyBits)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "sign/"+role, map[string]interface{}{
"common_name": "localhost",
"csr": csrPem,
})
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
haveErr := err != nil || resp == nil
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
if haveErr != test.ExpectError {
t.Fatalf("key size regression test [%d] failed: haveErr: %v, expectErr: %v, err: %v, resp: %v, test case: %v, caKeyType: %v, caKeyBits: %v, role: %v, keyType: %v, keyBits: %v", index, haveErr, test.ExpectError, err, resp, test, caKeyType, caKeyBits, role, keyType, keyBits)
}
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
Add PSS support to PKI Secrets Engine (#16519) * Add PSS signature support to Vault PKI engine Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use issuer's RevocationSigAlg for CRL signing We introduce a new parameter on issuers, revocation_signature_algorithm to control the signature algorithm used during CRL signing. This is because the SignatureAlgorithm value from the certificate itself is incorrect for this purpose: a RSA root could sign an ECDSA intermediate with say, SHA256WithRSA, but when the intermediate goes to sign a CRL, it must use ECDSAWithSHA256 or equivalent instead of SHA256WithRSA. When coupled with support for PSS-only keys, allowing the user to set the signature algorithm value as desired seems like the best approach. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add use_pss, revocation_signature_algorithm docs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add PSS to signature role issuance test matrix Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow roots to self-identify revocation alg When using PSS support with a managed key, sometimes the underlying device will not support PKCS#1v1.5 signatures. This results in CRL building failing, unless we update the entry's signature algorithm prior to building the CRL for the new root. With a RSA-type key and use_pss=true, we use the signature bits value to decide which hash function to use for PSS support. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add clearer error message on failed import When CRL building fails during cert/key import, due to PSS failures, give a better indication to the user that import succeeded its just CRL building that failed. This tells them the parameter to adjust on the issuer and warns that CRL building will fail until this is fixed. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add case insensitive SigAlgo matching Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert UsePSS back to regular bool Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor PSS->certTemplate into helper function Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Proper string output on rev_sig_alg display Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Copy root's SignatureAlgorithm for CRL building Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-03 16:42:24 +00:00
if resp != nil && test.RoleUsePSS && caKeyType == "rsa" {
leafCert := parseCert(t, resp.Data["certificate"].(string))
switch leafCert.SignatureAlgorithm {
case x509.SHA256WithRSAPSS, x509.SHA384WithRSAPSS, x509.SHA512WithRSAPSS:
default:
t.Fatalf("key size regression test [%d] failed on role %v: unexpected signature algorithm; expected RSA-type CA to sign a leaf cert with PSS algorithm; got %v", index, role, leafCert.SignatureAlgorithm.String())
}
}
tested += 1
}
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
}
}
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBDelete(b, s, "root")
if err != nil {
t.Fatal(err)
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
}
}
}
return tested
}
func TestBackend_Roles_KeySizeRegression(t *testing.T) {
t.Parallel()
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
// Regression testing of role's issuance policy.
testCases := []KeySizeRegression{
// RSA with default parameters should fail to issue smaller RSA keys
// and any size ECDSA/Ed25519 keys.
Add PSS support to PKI Secrets Engine (#16519) * Add PSS signature support to Vault PKI engine Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use issuer's RevocationSigAlg for CRL signing We introduce a new parameter on issuers, revocation_signature_algorithm to control the signature algorithm used during CRL signing. This is because the SignatureAlgorithm value from the certificate itself is incorrect for this purpose: a RSA root could sign an ECDSA intermediate with say, SHA256WithRSA, but when the intermediate goes to sign a CRL, it must use ECDSAWithSHA256 or equivalent instead of SHA256WithRSA. When coupled with support for PSS-only keys, allowing the user to set the signature algorithm value as desired seems like the best approach. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add use_pss, revocation_signature_algorithm docs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add PSS to signature role issuance test matrix Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow roots to self-identify revocation alg When using PSS support with a managed key, sometimes the underlying device will not support PKCS#1v1.5 signatures. This results in CRL building failing, unless we update the entry's signature algorithm prior to building the CRL for the new root. With a RSA-type key and use_pss=true, we use the signature bits value to decide which hash function to use for PSS support. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add clearer error message on failed import When CRL building fails during cert/key import, due to PSS failures, give a better indication to the user that import succeeded its just CRL building that failed. This tells them the parameter to adjust on the issuer and warns that CRL building will fail until this is fixed. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add case insensitive SigAlgo matching Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert UsePSS back to regular bool Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor PSS->certTemplate into helper function Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Proper string output on rev_sig_alg display Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Copy root's SignatureAlgorithm for CRL building Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-03 16:42:24 +00:00
/* 0 */ {"rsa", []int{0, 2048}, []int{0, 256, 384, 512}, false, []string{"rsa", "ec", "ec", "ec", "ec", "ed25519"}, []int{1024, 224, 256, 384, 521, 0}, true},
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
// But it should work to issue larger RSA keys.
Add PSS support to PKI Secrets Engine (#16519) * Add PSS signature support to Vault PKI engine Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use issuer's RevocationSigAlg for CRL signing We introduce a new parameter on issuers, revocation_signature_algorithm to control the signature algorithm used during CRL signing. This is because the SignatureAlgorithm value from the certificate itself is incorrect for this purpose: a RSA root could sign an ECDSA intermediate with say, SHA256WithRSA, but when the intermediate goes to sign a CRL, it must use ECDSAWithSHA256 or equivalent instead of SHA256WithRSA. When coupled with support for PSS-only keys, allowing the user to set the signature algorithm value as desired seems like the best approach. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add use_pss, revocation_signature_algorithm docs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add PSS to signature role issuance test matrix Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow roots to self-identify revocation alg When using PSS support with a managed key, sometimes the underlying device will not support PKCS#1v1.5 signatures. This results in CRL building failing, unless we update the entry's signature algorithm prior to building the CRL for the new root. With a RSA-type key and use_pss=true, we use the signature bits value to decide which hash function to use for PSS support. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add clearer error message on failed import When CRL building fails during cert/key import, due to PSS failures, give a better indication to the user that import succeeded its just CRL building that failed. This tells them the parameter to adjust on the issuer and warns that CRL building will fail until this is fixed. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add case insensitive SigAlgo matching Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert UsePSS back to regular bool Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor PSS->certTemplate into helper function Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Proper string output on rev_sig_alg display Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Copy root's SignatureAlgorithm for CRL building Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-03 16:42:24 +00:00
/* 1 */ {"rsa", []int{0, 2048}, []int{0, 256, 384, 512}, false, []string{"rsa", "rsa"}, []int{2048, 3072}, false},
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
// EC with default parameters should fail to issue smaller EC keys
// and any size RSA/Ed25519 keys.
Add PSS support to PKI Secrets Engine (#16519) * Add PSS signature support to Vault PKI engine Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use issuer's RevocationSigAlg for CRL signing We introduce a new parameter on issuers, revocation_signature_algorithm to control the signature algorithm used during CRL signing. This is because the SignatureAlgorithm value from the certificate itself is incorrect for this purpose: a RSA root could sign an ECDSA intermediate with say, SHA256WithRSA, but when the intermediate goes to sign a CRL, it must use ECDSAWithSHA256 or equivalent instead of SHA256WithRSA. When coupled with support for PSS-only keys, allowing the user to set the signature algorithm value as desired seems like the best approach. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add use_pss, revocation_signature_algorithm docs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add PSS to signature role issuance test matrix Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow roots to self-identify revocation alg When using PSS support with a managed key, sometimes the underlying device will not support PKCS#1v1.5 signatures. This results in CRL building failing, unless we update the entry's signature algorithm prior to building the CRL for the new root. With a RSA-type key and use_pss=true, we use the signature bits value to decide which hash function to use for PSS support. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add clearer error message on failed import When CRL building fails during cert/key import, due to PSS failures, give a better indication to the user that import succeeded its just CRL building that failed. This tells them the parameter to adjust on the issuer and warns that CRL building will fail until this is fixed. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add case insensitive SigAlgo matching Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert UsePSS back to regular bool Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor PSS->certTemplate into helper function Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Proper string output on rev_sig_alg display Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Copy root's SignatureAlgorithm for CRL building Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-03 16:42:24 +00:00
/* 2 */ {"ec", []int{0}, []int{0}, false, []string{"rsa", "ec", "ed25519"}, []int{2048, 224, 0}, true},
// But it should work to issue larger EC keys. Note that we should be
// independent of signature bits as that's computed from the issuer
// type (for EC based issuers).
Add PSS support to PKI Secrets Engine (#16519) * Add PSS signature support to Vault PKI engine Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use issuer's RevocationSigAlg for CRL signing We introduce a new parameter on issuers, revocation_signature_algorithm to control the signature algorithm used during CRL signing. This is because the SignatureAlgorithm value from the certificate itself is incorrect for this purpose: a RSA root could sign an ECDSA intermediate with say, SHA256WithRSA, but when the intermediate goes to sign a CRL, it must use ECDSAWithSHA256 or equivalent instead of SHA256WithRSA. When coupled with support for PSS-only keys, allowing the user to set the signature algorithm value as desired seems like the best approach. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add use_pss, revocation_signature_algorithm docs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add PSS to signature role issuance test matrix Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow roots to self-identify revocation alg When using PSS support with a managed key, sometimes the underlying device will not support PKCS#1v1.5 signatures. This results in CRL building failing, unless we update the entry's signature algorithm prior to building the CRL for the new root. With a RSA-type key and use_pss=true, we use the signature bits value to decide which hash function to use for PSS support. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add clearer error message on failed import When CRL building fails during cert/key import, due to PSS failures, give a better indication to the user that import succeeded its just CRL building that failed. This tells them the parameter to adjust on the issuer and warns that CRL building will fail until this is fixed. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add case insensitive SigAlgo matching Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert UsePSS back to regular bool Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor PSS->certTemplate into helper function Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Proper string output on rev_sig_alg display Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Copy root's SignatureAlgorithm for CRL building Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-03 16:42:24 +00:00
/* 3 */ {"ec", []int{224}, []int{0, 256, 384, 521}, false, []string{"ec", "ec", "ec", "ec"}, []int{224, 256, 384, 521}, false},
/* 4 */ {"ec", []int{0, 256}, []int{0, 256, 384, 521}, false, []string{"ec", "ec", "ec"}, []int{256, 384, 521}, false},
/* 5 */ {"ec", []int{384}, []int{0, 256, 384, 521}, false, []string{"ec", "ec"}, []int{384, 521}, false},
/* 6 */ {"ec", []int{521}, []int{0, 256, 384, 512}, false, []string{"ec"}, []int{521}, false},
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
// Ed25519 should reject RSA and EC keys.
Add PSS support to PKI Secrets Engine (#16519) * Add PSS signature support to Vault PKI engine Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use issuer's RevocationSigAlg for CRL signing We introduce a new parameter on issuers, revocation_signature_algorithm to control the signature algorithm used during CRL signing. This is because the SignatureAlgorithm value from the certificate itself is incorrect for this purpose: a RSA root could sign an ECDSA intermediate with say, SHA256WithRSA, but when the intermediate goes to sign a CRL, it must use ECDSAWithSHA256 or equivalent instead of SHA256WithRSA. When coupled with support for PSS-only keys, allowing the user to set the signature algorithm value as desired seems like the best approach. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add use_pss, revocation_signature_algorithm docs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add PSS to signature role issuance test matrix Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow roots to self-identify revocation alg When using PSS support with a managed key, sometimes the underlying device will not support PKCS#1v1.5 signatures. This results in CRL building failing, unless we update the entry's signature algorithm prior to building the CRL for the new root. With a RSA-type key and use_pss=true, we use the signature bits value to decide which hash function to use for PSS support. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add clearer error message on failed import When CRL building fails during cert/key import, due to PSS failures, give a better indication to the user that import succeeded its just CRL building that failed. This tells them the parameter to adjust on the issuer and warns that CRL building will fail until this is fixed. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add case insensitive SigAlgo matching Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert UsePSS back to regular bool Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor PSS->certTemplate into helper function Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Proper string output on rev_sig_alg display Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Copy root's SignatureAlgorithm for CRL building Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-03 16:42:24 +00:00
/* 7 */ {"ed25519", []int{0}, []int{0}, false, []string{"rsa", "ec", "ec"}, []int{2048, 256, 521}, true},
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
// But it should work to issue Ed25519 keys.
Add PSS support to PKI Secrets Engine (#16519) * Add PSS signature support to Vault PKI engine Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use issuer's RevocationSigAlg for CRL signing We introduce a new parameter on issuers, revocation_signature_algorithm to control the signature algorithm used during CRL signing. This is because the SignatureAlgorithm value from the certificate itself is incorrect for this purpose: a RSA root could sign an ECDSA intermediate with say, SHA256WithRSA, but when the intermediate goes to sign a CRL, it must use ECDSAWithSHA256 or equivalent instead of SHA256WithRSA. When coupled with support for PSS-only keys, allowing the user to set the signature algorithm value as desired seems like the best approach. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add use_pss, revocation_signature_algorithm docs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add PSS to signature role issuance test matrix Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow roots to self-identify revocation alg When using PSS support with a managed key, sometimes the underlying device will not support PKCS#1v1.5 signatures. This results in CRL building failing, unless we update the entry's signature algorithm prior to building the CRL for the new root. With a RSA-type key and use_pss=true, we use the signature bits value to decide which hash function to use for PSS support. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add clearer error message on failed import When CRL building fails during cert/key import, due to PSS failures, give a better indication to the user that import succeeded its just CRL building that failed. This tells them the parameter to adjust on the issuer and warns that CRL building will fail until this is fixed. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add case insensitive SigAlgo matching Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert UsePSS back to regular bool Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor PSS->certTemplate into helper function Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Proper string output on rev_sig_alg display Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Copy root's SignatureAlgorithm for CRL building Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-03 16:42:24 +00:00
/* 8 */ {"ed25519", []int{0}, []int{0}, false, []string{"ed25519"}, []int{0}, false},
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
// Any key type should reject insecure RSA key sizes.
Add PSS support to PKI Secrets Engine (#16519) * Add PSS signature support to Vault PKI engine Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use issuer's RevocationSigAlg for CRL signing We introduce a new parameter on issuers, revocation_signature_algorithm to control the signature algorithm used during CRL signing. This is because the SignatureAlgorithm value from the certificate itself is incorrect for this purpose: a RSA root could sign an ECDSA intermediate with say, SHA256WithRSA, but when the intermediate goes to sign a CRL, it must use ECDSAWithSHA256 or equivalent instead of SHA256WithRSA. When coupled with support for PSS-only keys, allowing the user to set the signature algorithm value as desired seems like the best approach. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add use_pss, revocation_signature_algorithm docs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add PSS to signature role issuance test matrix Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow roots to self-identify revocation alg When using PSS support with a managed key, sometimes the underlying device will not support PKCS#1v1.5 signatures. This results in CRL building failing, unless we update the entry's signature algorithm prior to building the CRL for the new root. With a RSA-type key and use_pss=true, we use the signature bits value to decide which hash function to use for PSS support. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add clearer error message on failed import When CRL building fails during cert/key import, due to PSS failures, give a better indication to the user that import succeeded its just CRL building that failed. This tells them the parameter to adjust on the issuer and warns that CRL building will fail until this is fixed. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add case insensitive SigAlgo matching Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert UsePSS back to regular bool Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor PSS->certTemplate into helper function Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Proper string output on rev_sig_alg display Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Copy root's SignatureAlgorithm for CRL building Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-03 16:42:24 +00:00
/* 9 */ {"any", []int{0}, []int{0, 256, 384, 512}, false, []string{"rsa", "rsa"}, []int{512, 1024}, true},
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
// But work for everything else.
Add PSS support to PKI Secrets Engine (#16519) * Add PSS signature support to Vault PKI engine Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use issuer's RevocationSigAlg for CRL signing We introduce a new parameter on issuers, revocation_signature_algorithm to control the signature algorithm used during CRL signing. This is because the SignatureAlgorithm value from the certificate itself is incorrect for this purpose: a RSA root could sign an ECDSA intermediate with say, SHA256WithRSA, but when the intermediate goes to sign a CRL, it must use ECDSAWithSHA256 or equivalent instead of SHA256WithRSA. When coupled with support for PSS-only keys, allowing the user to set the signature algorithm value as desired seems like the best approach. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add use_pss, revocation_signature_algorithm docs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add PSS to signature role issuance test matrix Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow roots to self-identify revocation alg When using PSS support with a managed key, sometimes the underlying device will not support PKCS#1v1.5 signatures. This results in CRL building failing, unless we update the entry's signature algorithm prior to building the CRL for the new root. With a RSA-type key and use_pss=true, we use the signature bits value to decide which hash function to use for PSS support. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add clearer error message on failed import When CRL building fails during cert/key import, due to PSS failures, give a better indication to the user that import succeeded its just CRL building that failed. This tells them the parameter to adjust on the issuer and warns that CRL building will fail until this is fixed. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add case insensitive SigAlgo matching Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert UsePSS back to regular bool Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor PSS->certTemplate into helper function Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Proper string output on rev_sig_alg display Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Copy root's SignatureAlgorithm for CRL building Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-03 16:42:24 +00:00
/* 10 */ {"any", []int{0}, []int{0, 256, 384, 512}, false, []string{"rsa", "rsa", "ec", "ec", "ec", "ec", "ed25519"}, []int{2048, 3072, 224, 256, 384, 521, 0}, false},
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
// RSA with larger than default key size should reject smaller ones.
Add PSS support to PKI Secrets Engine (#16519) * Add PSS signature support to Vault PKI engine Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use issuer's RevocationSigAlg for CRL signing We introduce a new parameter on issuers, revocation_signature_algorithm to control the signature algorithm used during CRL signing. This is because the SignatureAlgorithm value from the certificate itself is incorrect for this purpose: a RSA root could sign an ECDSA intermediate with say, SHA256WithRSA, but when the intermediate goes to sign a CRL, it must use ECDSAWithSHA256 or equivalent instead of SHA256WithRSA. When coupled with support for PSS-only keys, allowing the user to set the signature algorithm value as desired seems like the best approach. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add use_pss, revocation_signature_algorithm docs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add PSS to signature role issuance test matrix Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow roots to self-identify revocation alg When using PSS support with a managed key, sometimes the underlying device will not support PKCS#1v1.5 signatures. This results in CRL building failing, unless we update the entry's signature algorithm prior to building the CRL for the new root. With a RSA-type key and use_pss=true, we use the signature bits value to decide which hash function to use for PSS support. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add clearer error message on failed import When CRL building fails during cert/key import, due to PSS failures, give a better indication to the user that import succeeded its just CRL building that failed. This tells them the parameter to adjust on the issuer and warns that CRL building will fail until this is fixed. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add case insensitive SigAlgo matching Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert UsePSS back to regular bool Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor PSS->certTemplate into helper function Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Proper string output on rev_sig_alg display Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Copy root's SignatureAlgorithm for CRL building Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-03 16:42:24 +00:00
/* 11 */ {"rsa", []int{3072}, []int{0, 256, 384, 512}, false, []string{"rsa"}, []int{2048}, true},
// We should be able to sign with PSS with any CA key type.
/* 12 */ {"rsa", []int{0}, []int{0, 256, 384, 512}, true, []string{"rsa"}, []int{2048}, false},
/* 13 */ {"ec", []int{0}, []int{0}, true, []string{"ec"}, []int{256}, false},
/* 14 */ {"ed25519", []int{0}, []int{0}, true, []string{"ed25519"}, []int{0}, false},
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
}
Add PSS support to PKI Secrets Engine (#16519) * Add PSS signature support to Vault PKI engine Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use issuer's RevocationSigAlg for CRL signing We introduce a new parameter on issuers, revocation_signature_algorithm to control the signature algorithm used during CRL signing. This is because the SignatureAlgorithm value from the certificate itself is incorrect for this purpose: a RSA root could sign an ECDSA intermediate with say, SHA256WithRSA, but when the intermediate goes to sign a CRL, it must use ECDSAWithSHA256 or equivalent instead of SHA256WithRSA. When coupled with support for PSS-only keys, allowing the user to set the signature algorithm value as desired seems like the best approach. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add use_pss, revocation_signature_algorithm docs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add PSS to signature role issuance test matrix Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow roots to self-identify revocation alg When using PSS support with a managed key, sometimes the underlying device will not support PKCS#1v1.5 signatures. This results in CRL building failing, unless we update the entry's signature algorithm prior to building the CRL for the new root. With a RSA-type key and use_pss=true, we use the signature bits value to decide which hash function to use for PSS support. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add clearer error message on failed import When CRL building fails during cert/key import, due to PSS failures, give a better indication to the user that import succeeded its just CRL building that failed. This tells them the parameter to adjust on the issuer and warns that CRL building will fail until this is fixed. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add case insensitive SigAlgo matching Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert UsePSS back to regular bool Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor PSS->certTemplate into helper function Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Proper string output on rev_sig_alg display Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Copy root's SignatureAlgorithm for CRL building Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-03 16:42:24 +00:00
if len(testCases) != 15 {
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
t.Fatalf("misnumbered test case entries will make it hard to find bugs: %v", len(testCases))
}
b, s := CreateBackendWithStorage(t)
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
tested := 0
for index, test := range testCases {
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
tested += RoleKeySizeRegressionHelper(t, b, s, index, test)
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
}
t.Logf("Key size regression expanded matrix test scenarios: %d", tested)
Fix handling of default zero SignatureBits value with Any key type in PKI Secrets Engine (#14875) * Correctly handle minimums, default SignatureBits When using KeyType = "any" on a role (whether explicitly or implicitly via a sign-verbatim like operation), we need to update the value of SignatureBits from its new value 0 to a per-key-type default value. This will allow sign operations on these paths to function correctly, having the correctly inferred default signature bit length. Additionally, this allows the computed default value for key type to be used for minimum size validation in the RSA/ECDSA paths. We additionally enforce the 2048-minimum in this case as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix defaults and validation of "any" KeyType When certutil is given the placeholder any keytype, it attempts to validate and update the default zero value. However, in lacking a default value for SignatureBits, it cannot update the value from the zero value, thus causing validation to fail. Add more awareness to the placeholder "any" value to certutil. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add role-based regression tests for key bits This adds regression tests for Key Type, Key Bits, and Signature Bits parameters on the role. We test several values, including the "any" value to ensure it correctly restricts key sizes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add sign-verbatim test for key type This ensures that we test sign-verbatim against a variety of key types. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Steven Clark <steven.clark@hashicorp.com>
2022-04-04 19:26:54 +00:00
}
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
func TestRootWithExistingKey(t *testing.T) {
t.Parallel()
b, s := CreateBackendWithStorage(t)
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
var err error
// Fail requests if type is existing, and we specify the key_type param
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "root/generate/existing", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "root myvault.com",
"key_type": "rsa",
})
require.Error(t, err)
require.Contains(t, err.Error(), "key_type nor key_bits arguments can be set in this mode")
// Fail requests if type is existing, and we specify the key_bits param
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "root/generate/existing", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "root myvault.com",
"key_bits": "2048",
})
require.Error(t, err)
require.Contains(t, err.Error(), "key_type nor key_bits arguments can be set in this mode")
// Fail if the specified key does not exist.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "issuers/generate/root/existing", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "root myvault.com",
"issuer_name": "my-issuer1",
"key_ref": "my-key1",
})
require.Error(t, err)
require.Contains(t, err.Error(), "unable to find PKI key for reference: my-key1")
// Fail if the specified key name is default.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "issuers/generate/root/internal", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "root myvault.com",
"issuer_name": "my-issuer1",
"key_name": "Default",
})
require.Error(t, err)
require.Contains(t, err.Error(), "reserved keyword 'default' can not be used as key name")
// Fail if the specified issuer name is default.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "issuers/generate/root/internal", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "root myvault.com",
"issuer_name": "DEFAULT",
})
require.Error(t, err)
require.Contains(t, err.Error(), "reserved keyword 'default' can not be used as issuer name")
// Create the first CA
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err := CBWrite(b, s, "issuers/generate/root/internal", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "root myvault.com",
"key_type": "rsa",
"issuer_name": "my-issuer1",
})
require.NoError(t, err)
require.NotNil(t, resp.Data["certificate"])
myIssuerId1 := resp.Data["issuer_id"]
myKeyId1 := resp.Data["key_id"]
require.NotEmpty(t, myIssuerId1)
require.NotEmpty(t, myKeyId1)
// Fetch the parsed CRL; it should be empty as we've not revoked anything
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
parsedCrl := getParsedCrlFromBackend(t, b, s, "issuer/my-issuer1/crl/der")
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
require.Equal(t, len(parsedCrl.TBSCertList.RevokedCertificates), 0, "should have no revoked certificates")
// Fail if the specified issuer name is re-used.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "issuers/generate/root/internal", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "root myvault.com",
"issuer_name": "my-issuer1",
})
require.Error(t, err)
require.Contains(t, err.Error(), "issuer name already in use")
// Create the second CA
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "issuers/generate/root/internal", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "root myvault.com",
"key_type": "rsa",
"issuer_name": "my-issuer2",
"key_name": "root-key2",
})
require.NoError(t, err)
require.NotNil(t, resp.Data["certificate"])
myIssuerId2 := resp.Data["issuer_id"]
myKeyId2 := resp.Data["key_id"]
require.NotEmpty(t, myIssuerId2)
require.NotEmpty(t, myKeyId2)
// Fetch the parsed CRL; it should be empty as we've not revoked anything
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
parsedCrl = getParsedCrlFromBackend(t, b, s, "issuer/my-issuer2/crl/der")
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
require.Equal(t, len(parsedCrl.TBSCertList.RevokedCertificates), 0, "should have no revoked certificates")
// Fail if the specified key name is re-used.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "issuers/generate/root/internal", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "root myvault.com",
"issuer_name": "my-issuer3",
"key_name": "root-key2",
})
require.Error(t, err)
require.Contains(t, err.Error(), "key name already in use")
// Create a third CA re-using key from CA 1
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "issuers/generate/root/existing", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "root myvault.com",
"issuer_name": "my-issuer3",
"key_ref": myKeyId1,
})
require.NoError(t, err)
require.NotNil(t, resp.Data["certificate"])
myIssuerId3 := resp.Data["issuer_id"]
myKeyId3 := resp.Data["key_id"]
require.NotEmpty(t, myIssuerId3)
require.NotEmpty(t, myKeyId3)
// Fetch the parsed CRL; it should be empty as we've not revoking anything.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
parsedCrl = getParsedCrlFromBackend(t, b, s, "issuer/my-issuer3/crl/der")
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
require.Equal(t, len(parsedCrl.TBSCertList.RevokedCertificates), 0, "should have no revoked certificates")
// Signatures should be the same since this is just a reissued cert. We
// use signature as a proxy for "these two CRLs are equal".
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
firstCrl := getParsedCrlFromBackend(t, b, s, "issuer/my-issuer1/crl/der")
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
require.Equal(t, parsedCrl.SignatureValue, firstCrl.SignatureValue)
require.NotEqual(t, myIssuerId1, myIssuerId2)
require.NotEqual(t, myIssuerId1, myIssuerId3)
require.NotEqual(t, myKeyId1, myKeyId2)
require.Equal(t, myKeyId1, myKeyId3)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBList(b, s, "issuers")
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
require.NoError(t, err)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
require.Equal(t, 3, len(resp.Data["keys"].([]string)))
require.Contains(t, resp.Data["keys"], string(myIssuerId1.(issuerID)))
require.Contains(t, resp.Data["keys"], string(myIssuerId2.(issuerID)))
require.Contains(t, resp.Data["keys"], string(myIssuerId3.(issuerID)))
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
}
func TestIntermediateWithExistingKey(t *testing.T) {
t.Parallel()
b, s := CreateBackendWithStorage(t)
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
var err error
// Fail requests if type is existing, and we specify the key_type param
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "intermediate/generate/existing", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "root myvault.com",
"key_type": "rsa",
})
require.Error(t, err)
require.Contains(t, err.Error(), "key_type nor key_bits arguments can be set in this mode")
// Fail requests if type is existing, and we specify the key_bits param
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "intermediate/generate/existing", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "root myvault.com",
"key_bits": "2048",
})
require.Error(t, err)
require.Contains(t, err.Error(), "key_type nor key_bits arguments can be set in this mode")
// Fail if the specified key does not exist.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "issuers/generate/intermediate/existing", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "root myvault.com",
"key_ref": "my-key1",
})
require.Error(t, err)
require.Contains(t, err.Error(), "unable to find PKI key for reference: my-key1")
// Create the first intermediate CA
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err := CBWrite(b, s, "issuers/generate/intermediate/internal", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "root myvault.com",
"key_type": "rsa",
})
require.NoError(t, err)
// csr1 := resp.Data["csr"]
myKeyId1 := resp.Data["key_id"]
require.NotEmpty(t, myKeyId1)
// Create the second intermediate CA
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "issuers/generate/intermediate/internal", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "root myvault.com",
"key_type": "rsa",
"key_name": "interkey1",
})
require.NoError(t, err)
// csr2 := resp.Data["csr"]
myKeyId2 := resp.Data["key_id"]
require.NotEmpty(t, myKeyId2)
// Create a third intermediate CA re-using key from intermediate CA 1
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "issuers/generate/intermediate/existing", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "root myvault.com",
"key_ref": myKeyId1,
})
require.NoError(t, err)
// csr3 := resp.Data["csr"]
myKeyId3 := resp.Data["key_id"]
require.NotEmpty(t, myKeyId3)
require.NotEqual(t, myKeyId1, myKeyId2)
require.Equal(t, myKeyId1, myKeyId3, "our new ca did not seem to reuse the key as we expected.")
}
func TestIssuanceTTLs(t *testing.T) {
t.Parallel()
b, s := CreateBackendWithStorage(t)
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err := CBWrite(b, s, "root/generate/internal", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "root example.com",
"issuer_name": "root",
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
"ttl": "10s",
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"key_type": "ec",
})
require.NoError(t, err)
require.NotNil(t, resp)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
rootCert := parseCert(t, resp.Data["certificate"].(string))
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "roles/local-testing", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"allow_any_name": true,
"enforce_hostnames": false,
"key_type": "ec",
})
require.NoError(t, err)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "issue/local-testing", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "testing",
"ttl": "1s",
})
require.NoError(t, err, "expected issuance to succeed due to shorter ttl than cert ttl")
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "issue/local-testing", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "testing",
})
require.Error(t, err, "expected issuance to fail due to longer default ttl than cert ttl")
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "issuer/root", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"issuer_name": "root",
"leaf_not_after_behavior": "permit",
})
require.NoError(t, err)
require.NotNil(t, resp)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "issue/local-testing", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "testing",
})
require.NoError(t, err, "expected issuance to succeed due to permitted longer TTL")
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "issuer/root", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"issuer_name": "root",
"leaf_not_after_behavior": "truncate",
})
require.NoError(t, err)
require.NotNil(t, resp)
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "issue/local-testing", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "testing",
})
require.NoError(t, err, "expected issuance to succeed due to truncated ttl")
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
// Sleep until the parent cert expires and the clock rolls over
// to the next second.
time.Sleep(time.Until(rootCert.NotAfter) + (1500 * time.Millisecond))
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
resp, err = CBWrite(b, s, "issuer/root", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"issuer_name": "root",
"leaf_not_after_behavior": "err",
})
require.NoError(t, err)
require.NotNil(t, resp)
// Even 1s ttl should now fail.
Refactor PKI tests for speed (#15999) * Refactor role issuance tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.879s After: github.com/hashicorp/vault/builtin/logical/pki 1.063s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor role key bit tests to use direct backend Also removes redundant cases. Before: github.com/hashicorp/vault/builtin/logical/pki 136.605s After: github.com/hashicorp/vault/builtin/logical/pki 24.713s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor common name test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.767s After: github.com/hashicorp/vault/builtin/logical/pki 0.611s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor device cert tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.725s After: github.com/hashicorp/vault/builtin/logical/pki 0.402s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor invalid parameter test to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 3.777s After: github.com/hashicorp/vault/builtin/logical/pki 0.021s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Alt Issuer tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.560s After: github.com/hashicorp/vault/builtin/logical/pki 0.111s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor root idempotency tests to use direct backend As a result, we've had to import a root cert from elsewhere in the test suite, rather than using the one off the cluster. Before: github.com/hashicorp/vault/builtin/logical/pki 4.399s After: github.com/hashicorp/vault/builtin/logical/pki 0.523s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Move PKI direct backend helpers to common location Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor OID SANs test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.284s After: github.com/hashicorp/vault/builtin/logical/pki 0.808s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor allowed serial numbers test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.789s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor URI SANs to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.245s After: github.com/hashicorp/vault/builtin/logical/pki 0.600s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Full Chain CA tests to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 14.503s After: github.com/hashicorp/vault/builtin/logical/pki 2.082s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update Allow Past CA tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.323s After: github.com/hashicorp/vault/builtin/logical/pki 0.322s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Convert existing-key root test to direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.430s After: github.com/hashicorp/vault/builtin/logical/pki 0.370s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor CRL enable/disable tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 5.738s After: github.com/hashicorp/vault/builtin/logical/pki 2.482s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update intermediate existing key tests to use direct backend Before: github.com/hashicorp/vault/builtin/logical/pki 4.182s After: github.com/hashicorp/vault/builtin/logical/pki 0.416s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor Issuance TTL verification tests to use direct backend Also shorten sleep duration slightly by precisely calculating it relative to the actual cert life time. Before: github.com/hashicorp/vault/builtin/logical/pki 19.755s After: github.com/hashicorp/vault/builtin/logical/pki 11.521s Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-06-16 13:11:22 +00:00
_, err = CBWrite(b, s, "issue/local-testing", map[string]interface{}{
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
"common_name": "testing",
"ttl": "1s",
})
require.Error(t, err, "expected issuance to fail due to longer default ttl than cert ttl")
}
func TestSealWrappedStorageConfigured(t *testing.T) {
t.Parallel()
b, _ := CreateBackendWithStorage(t)
Allow Multiple Issuers in PKI Secret Engine Mounts - PKI Pod (#15277) * Starter PKI CA Storage API (#14796) * Simple starting PKI storage api for CA rotation * Add key and issuer storage apis * Add listKeys and listIssuers storage implementations * Add simple keys and issuers configuration storage api methods * Handle resolving key, issuer references The API context will usually have a user-specified reference to the key. This is either the literal string "default" to select the default key, an identifier of the key, or a slug name for the key. Here, we wish to resolve this reference to an actual identifier that can be understood by storage. Also adds the missing Name field to keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add method to fetch an issuer's cert bundle This adds a method to construct a certutil.CertBundle from the specified issuer identifier, optionally loading its corresponding key for signing. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor certutil PrivateKey PEM handling This refactors the parsing of PrivateKeys from PEM blobs into shared methods (ParsePEMKey, ParseDERKey) that can be reused by the existing Bundle parsing logic (ParsePEMBundle) or independently in the new issuers/key-based PKI storage code. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add importKey, importCert to PKI storage importKey is generally preferable to the low-level writeKey for adding new entries. This takes only the contents of the private key (as a string -- so a PEM bundle or a managed key handle) and checks if it already exists in the storage. If it does, it returns the existing key instance. Otherwise, we create a new one. In the process, we detect any issuers using this key and link them back to the new key entry. The same holds for importCert over importKey, with the note that keys are not modified when importing certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for importing issuers, keys This adds tests for importing keys and issuers into the new storage layout, ensuring that identifiers are correctly inferred and linked. Note that directly writing entries to storage (writeKey/writeissuer) will take KeyID links from the parent entry and should not be used for import; only existing entries should be updated with this info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Implement PKI storage migration. - Hook into the backend::initialize function, calling the migration on a primary only. - Migrate an existing certificate bundle to the new issuers and key layout * Make fetchCAInfo aware of new storage layout This allows fetchCAInfo to fetch a specified issuer, via a reference parameter provided by the user. We pass that into the storage layer and have it return a cert bundle for us. Finally, we need to validate that it truly has the key desired. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Begin /issuers API endpoints This implements the fetch operations around issuers in the PKI Secrets Engine. We implement the following operations: - LIST /issuers - returns a list of known issuers' IDs and names. - GET /issuer/:ref - returns a JSON blob with information about this issuer. - POST /issuer/:ref - allows configuring information about issuers, presently just its name. - DELETE /issuer/:ref - allows deleting the specified issuer. - GET /issuer/:ref/{der,pem} - returns a raw API response with just the DER (or PEM) of the issuer's certificate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add import to PKI Issuers API This adds the two core import code paths to the API: /issuers/import/cert and /issuers/import/bundle. The former differs from the latter in that the latter allows the import of keys. This allows operators to restrict importing of keys to privileged roles, while allowing more operators permission to import additional certificates (not used for signing, but instead for path/chain building). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-intermediate endpoint This endpoint allows existing issuers to be used to sign intermediate CA certificates. In the process, we've updated the existing /root/sign-intermediate endpoint to be equivalent to a call to /issuer/default/sign-intermediate. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-self-issued endpoint This endpoint allows existing issuers to be used to sign self-signed certificates. In the process, we've updated the existing /root/sign-self-issued endpoint to be equivalent to a call to /issuer/default/sign-self-issued. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/sign-verbatim endpoint This endpoint allows existing issuers to be used to directly sign CSRs. In the process, we've updated the existing /sign-verbatim endpoint to be equivalent to a call to /issuer/:ref/sign-verbatim. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow configuration of default issuers Using the new updateDefaultIssuerId(...) from the storage migration PR allows for easy implementation of configuring the default issuer. We restrict callers from setting blank defaults and setting default to default. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix fetching default issuers After setting a default issuer, one should be able to use the old /ca, /ca_chain, and /cert/{ca,ca_chain} endpoints to fetch the default issuer (and its chain). Update the fetchCertBySerial helper to no longer support fetching the ca and prefer fetchCAInfo for that instead (as we've already updated that to support fetching the new issuer location). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add /issuer/:ref/{sign,issue}/:role This updates the /sign and /issue endpoints, allowing them to take the default issuer (if none is provided by a role) and adding issuer-specific versions of them. Note that at this point in time, the behavior isn't yet ideal (as /sign/:role allows adding the ref=... parameter to override the default issuer); a later change adding role-based issuer specification will fix this incorrect behavior. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support root issuer generation * Add support for issuer generate intermediate end-point * Update issuer and key arguments to consistent values - Update all new API endpoints to use the new agreed upon argument names. - issuer_ref & key_ref to refer to existing - issuer_name & key_name for new definitions - Update returned values to always user issuer_id and key_id * Add utility methods to fetch common ref and name arguments - Add utility methods to fetch the issuer_name, issuer_ref, key_name and key_ref arguments from data fields. - Centralize the logic to clean up these inputs and apply various validations to all of them. * Rename common PKI backend handlers - Use the buildPath convention for the function name instead of common... * Move setting PKI defaults from writeCaBundle to proper import{keys,issuer} methods - PR feedback, move setting up the default configuration references within the import methods instead of within the writeCaBundle method. This should now cover all use cases of us setting up the defaults properly. * Introduce constants for issuer_ref, rename isKeyDefaultSet... * Fix legacy PKI sign-verbatim api path - Addresses some test failures due to an incorrect refactoring of a legacy api path /sign-verbatim within PKI * Use import code to handle intermediate, config/ca The existing bundle import code will satisfy the intermediate import; use it instead of the old ca_bundle import logic. Additionally, update /config/ca to use the new import code as well. While testing, a panic was discovered: > reflect.Value.SetMapIndex: value of type string is not assignable to type pki.keyId This was caused by returning a map with type issuerId->keyId; instead switch to returning string->string maps so the audit log can properly HMAC them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error message on missing defaults When the default issuer and key are missing (and haven't yet been specified), we should clarify that error message. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update test semantics for new changes This makes two minor changes to the existing test suite: 1. Importing partial bundles should now succeed, where they'd previously error. 2. fetchCertBySerial no longer handles CA certificates. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add support for deleting all keys, issuers The old DELETE /root code must now delete all keys and issuers for backwards compatibility. We strongly suggest calling individual delete methods (DELETE /key/:key_ref or DELETE /issuer/:issuer_ref) instead, for finer control. In the process, we detect whether the deleted key/issuers was set as the default. This will allow us to warn (from the single key/deletion issuer code) whether or not the default was deleted (while allowing the operation to succeed). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Introduce defaultRef constant within PKI - Replace hardcoded "default" references with a constant to easily identify various usages. - Use the addIssuerRefField function instead of redefining the field in various locations. * Rework PKI test TestBackend_Root_Idempotency - Validate that generate/root calls are no longer idempotent, but the bundle importing does not generate new keys/issuers - As before make sure that the delete root api resets everything - Address a bug within the storage that we bombed when we had multiple different key types within storage. * Assign Name=current to migrated key and issuer - Detail I missed from the RFC was to assign the Name field as "current" for migrated key and issuer. * Build CRL upon PKI intermediary set-signed api called - Add a call to buildCRL if we created an issuer within pathImportIssuers - Augment existing FullCAChain to verify we have a proper CRL post set-signed api call - Remove a code block writing out "ca" storage entry that is no longer used. * Identify which certificate or key failed When importing complex chains, we should identify in which certificate or key the failure occurred. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI migration writes out empty migration log entry - Since the elements of the struct were not exported we serialized an empty migration log to disk and would re-run the migration * Add chain-building logic to PKI issuers path With the one-entry-per-issuer approach, CA Chains become implicitly constructed from the pool of issuers. This roughly matches the existing expectations from /config/ca (wherein a chain could be provided) and /intemediate/set-signed (where a chain may be provided). However, in both of those cases, we simply accepted a chain. Here, we need to be able to reconstruct the chain from parts on disk. However, with potential rotation of roots, we need to be aware of disparate chains. Simply concating together all issuers isn't sufficient. Thus we need to be able to parse a certificate's Issuer and Subject field and reconstruct valid (and potentially parallel) parent<->child mappings. This attempts to handle roots, intermediates, cross-signed intermediates, cross-signed roots, and rotated keys (wherein one might not have a valid signature due to changed key material with the same subject). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Return CA Chain when fetching issuers This returns the CA Chain attribute of an issuer, showing its computed chain based on other issuers in the database, when fetching a specific issuer. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add testing for chain building Using the issuance infrastructure, we generate new certificates (either roots or intermediates), positing that this is roughly equivalent to importing an external bundle (minus error handling during partial imports). This allows us to incrementally construct complex chains, creating reissuance cliques and cross-signing cycles. By using ECDSA certificates, we avoid high signature verification and key generation times. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow manual construction of issuer chain Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix handling of duplicate names With the new issuer field (manual_chain), we can no longer err when a name already exists: we might be updating the existing issuer (with the same name), but changing its manual_chain field. Detect this error and correctly handle it. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for manual chain building We break the clique, instead building these chains manually, ensuring that the remaining chains do not change and only the modified certs change. We then reset them (back to implicit chain building) and ensure we get the same results as earlier. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter verification of issuers PEM format This ensures each issuer is only a single certificate entry (as validated by count and parsing) without any trailing data. We further ensure that each certificate PEM has leading and trailing spaces removed with only a single trailing new line remaining. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix full chain building Don't set the legacy IssuingCA field on the certificate bundle, as we prefer the CAChain field over it. Additionally, building the full chain could result in duplicate certificates when the CAChain included the leaf certificate itself. When building the full chain, ensure we don't include the bundle's certificate twice. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add stricter tests for full chain construction We wish to ensure that each desired certificate in the chain is only present once. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename PKI types to avoid constant variable name collisions keyId -> keyID issuerId -> issuerID key -> keyEntry issuer -> issuerEntry keyConfig -> keyConfigEntry issuerConfig -> issuerConfigEntry * Update CRL handling for multiple issuers When building CRLs, we've gotta make sure certs issued by that issuer land up on that issuer's CRL and not some other CRL. If no CRL is found (matching a cert), we'll place it on the default CRL. However, in the event of equivalent issuers (those with the same subject AND the same key material) -- perhaps due to reissuance -- we'll only create a single (unified) CRL for them. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching updated CRL locations This updates fetchCertBySerial to support querying the default issuer's CRL. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL storage location test case Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update to CRLv2 Format to copy RawIssuer When using the older Certificate.CreateCRL(...) call, Go's x509 library copies the parsed pkix.Name version of the CRL Issuer's Subject field. For certain constructed CAs, this fails since pkix.Name is not suitable for round-tripping. This also builds a CRLv1 (per RFC 5280) CRL. In updating to the newer x509.CreateRevocationList(...) call, we can construct the CRL in the CRLv2 format and correctly copy the issuer's name. However, this requires holding an additional field per-CRL, the CRLNumber field, which is required in Go's implementation of CRLv2 (though OPTIONAL in the spec). We store this on the new LocalCRLConfigEntry object, per-CRL. Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add comment regarding CRL non-assignment in GOTO In previous versions of Vault, it was possible to sign an empty CRL (when the CRL was disabled and a force-rebuild was requested). Add a comment about this case. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow fetching the specified issuer's CRL We add a new API endpoint to fetch the specified issuer's CRL directly (rather than the default issuer's CRL at /crl and /certs/crl). We also add a new test to validate the CRL in a multi-root scenario and ensure it is signed with the correct keys. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add new PKI key prefix to seal wrapped storage (#15126) * Refactor common backend initialization within backend_test - Leverage an existing helper method within the PKI backend tests to setup a PKI backend with storage. * Add ability to read legacy cert bundle if the migration has not occurred on secondaries. - Track the migration state forbidding an issuer/key writing api call if we have not migrated - For operations that just need to read the CA bundle, use the same tracking variable to switch between reading the legacy bundle or use the new key/issuer storage. - Add an invalidation function that will listen for updates to our log path to refresh the state on secondary clusters. * Always write migration entry to trigger secondary clusters to wake up - Some PR feedback and handle a case in which the primary cluster does not have a CA bundle within storage but somehow a secondary does. * Update CA Chain to report entire chain This merges the ca_chain JSON field (of the /certs/ca_chain path) with the regular certificate field, returning the root of trust always. This also affects the non-JSON (raw) endpoints as well. We return the default issuer's chain here, rather than all known issuers (as that may not form a strict chain). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow explicit issuer override on roles When a role is used to generate a certificate (such as with the sign/ and issue/ legacy paths or the legacy sign-verbatim/ paths), we prefer that issuer to the one on the request. This allows operators to set an issuer (other than default) for requests to be issued against, effectively making the change no different from the users' perspective as it is "just" a different role name. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for role-based issuer selection Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Expand NotAfter limit enforcement behavior Vault previously strictly enforced NotAfter/ttl values on certificate requests, erring if the requested TTL extended past the NotAfter date of the issuer. In the event of issuing an intermediate, this behavior was ignored, instead permitting the issuance. Users generally do not think to check their issuer's NotAfter date when requesting a certificate; thus this behavior was generally surprising. Per RFC 5280 however, issuers need to maintain status information throughout the life cycle of the issued cert. If this leaf cert were to be issued for a longer duration than the parent issuer, the CA must still maintain revocation information past its expiration. Thus, we add an option to the issuer to change the desired behavior: - err, to err out, - permit, to permit the longer NotAfter date, or - truncate, to silently truncate the expiration to the issuer's NotAfter date. Since expiration of certificates in the system's trust store are not generally validated (when validating an arbitrary leaf, e.g., during TLS validation), permit should generally only be used in that case. However, browsers usually validate intermediate's validity periods, and thus truncate should likely be used (as with permit, the leaf's chain will not validate towards the end of the issuance period). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for expanded issuance behaviors Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add warning on keyless default issuer (#15178) Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Update PKI to new Operations framework (#15180) The backend Framework has updated Callbacks (used extensively in PKI) to become deprecated; Operations takes their place and clarifies forwarding of requests. We switch to the new format everywhere, updating some bad assumptions about forwarding along the way. Anywhere writes are handled (that should be propagated to all nodes in all clusters), we choose to forward the request all the way up to the performance primary cluster's primary node. This holds for issuers/keys, roles, and configs (such as CRL config, which is globally set for all clusters despite all clusters having their own separate CRL). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Kitography/vault 5474 rebase (#15150) * These parts work (put in signature so that backend wouldn't break, but missing fields, desc, etc.) * Import and Generate API calls w/ needed additions to SDK. * make fmt * Add Help/Sync Text, fix some of internal/exported/kms code. * Fix PEM/DER Encoding issue. * make fmt * Standardize keyIdParam, keyNameParam, keyTypeParam * Add error response if key to be deleted is in use. * replaces all instances of "default" in code with defaultRef * Updates from Callbacks to Operations Function with explicit forwarding. * Fixes a panic with names not being updated everywhere. * add a logged error in addition to warning on deleting default key. * Normalize whitespace upon importing keys. Authored-by: Alexander Scheel <alexander.m.scheel@gmail.com> * Fix isKeyInUse functionality. * Fixes tests associated with newline at end of key pem. * Add alternative proposal PKI aliased paths (#15211) * Add aliased path for root/rotate/:exported This adds a user-friendly path name for generating a rotated root. We automatically choose the name "next" for the newly generated root at this path if it doesn't already exist. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add aliased path for intermediate/cross-sign This allows cross-signatures to work. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add path for replacing the current root This updates default to point to the value of the issuer with name "next" rather than its current value. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove plural issuers/ in signing paths These paths use a single issuer and thus shouldn't include the plural issuers/ as a path prefix, instead using the singular issuer/ path prefix. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Only warn if default issuer was imported When the default issuer was not (re-)imported, we'd fail to find it, causing an extraneous warning about missing keys, even though this issuer indeed had a key. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing issuer sign/issue paths Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clean up various warnings within the PKI package (#15230) * Rebuild CRLs on secondary performance clusters post migration and on new/updated issuers - Hook into the backend invalidation function so that secondaries are notified of new/updated issuer or migrations occuring on the primary cluster. Upon notification schedule a CRL rebuild to take place upon the next process to read/update the CRL or within the periodic function if no request comes in. * Schedule rebuilding PKI CRLs on active nodes only - Address an issue that we were scheduling the rebuilding of a CRL on standby nodes, which would not be able to write to storage. - Fix an issue with standby nodes not correctly determining that a migration previously occurred. * Return legacy CRL storage path when no migration has occurred. * Handle issuer, keys locking (#15227) * Handle locking of issuers during writes We need a write lock around writes to ensure serialization of modifications. We use a single lock for both issuer and key updates, in part because certain operations (like deletion) will potentially affect both. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add missing b.useLegacyBundleCaStorage guards Several locations needed to guard against early usage of the new issuers endpoint pre-migration. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Address PKI to properly support managed keys (#15256) * Address codebase for managed key fixes * Add proper public key comparison for better managed key support to importKeys * Remove redundant public key fetching within PKI importKeys * Correctly handle rebuilding remaining chains When deleting a specific issuer, we might impact the chains. From a consistency perspective, we need to ensure the remaining chains are correct and don't refer to the since-deleted issuer, so trigger a full rebuild here. We don't need to call this in the delete-the-world (DELETE /root) code path, as there shouldn't be any remaining issuers or chains to build. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove legacy CRL bundle on world deletion When calling DELETE /root, we should remove the legacy CRL bundle, since we're deleting the legacy CA issuer bundle as well. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Remove deleted issuers' CRL entries Since CRLs are no longer resolvable after deletion (due to missing issuer ID, which will cause resolution to fail regardless of if an ID or a name/default reference was used), we should delete these CRLs from storage to avoid leaking them. In the event that this issuer comes back (with key material), we can simply rebuild the CRL at that time (from the remaining revoked storage entries). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthed JSON fetching of CRLs, Issuers (#15253) Default to fetching JSON CRL for consistency This makes the bare issuer-specific CRL fetching endpoint return the JSON-wrapped CRL by default, moving the DER CRL to a specific endpoint. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add JSON-specific endpoint for fetching issuers Unlike the unqualified /issuer/:ref endpoint (which also returns JSON), we have a separate /issuer/:ref/json endpoint to return _only_ the PEM-encoded certificate and the chain, mirroring the existing /cert/ca endpoint but for a specific issuer. This allows us to make the endpoint unauthenticated, whereas the bare endpoint would remain authenticated and usually privileged. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Add tests for raw JSON endpoints Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add unauthenticated issuers endpoints to PKI table This adds the unauthenticated issuers endpoints? - LIST /issuers, - Fetching _just_ the issuer certificates (in JSON/DER/PEM form), and - Fetching the CRL of this issuer (in JSON/DER/PEM form). Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add issuer usage restrictions bitset This allows issuers to have usage restrictions, limiting whether they can be used to issue certificates or if they can generate CRLs. This allows certain issuers to not generate a CRL (if the global config is with the CRL enabled) or allows the issuer to not issue new certificates (but potentially letting the CRL generation continue). Setting both fields to false effectively forms a soft delete capability. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * PKI Pod rotation Add Base Changelog (#15283) * PKI Pod rotation changelog. * Use feature release-note formatting of changelog. Co-authored-by: Steven Clark <steven.clark@hashicorp.com> Co-authored-by: Kit Haines <kit.haines@hashicorp.com> Co-authored-by: kitography <khaines@mit.edu>
2022-05-11 16:42:28 +00:00
wrappedEntries := b.Backend.PathsSpecial.SealWrapStorage
// Make sure our legacy bundle is within the list
// NOTE: do not convert these test values to constants, we should always have these paths within seal wrap config
require.Contains(t, wrappedEntries, "config/ca_bundle", "Legacy bundle missing from seal wrap")
// The trailing / is important as it treats the entire folder requiring seal wrapping, not just config/key
require.Contains(t, wrappedEntries, "config/key/", "key prefix with trailing / missing from seal wrap.")
}
func TestBackend_ConfigCA_WithECParams(t *testing.T) {
t.Parallel()
b, s := CreateBackendWithStorage(t)
// Generated key with OpenSSL:
// $ openssl ecparam -out p256.key -name prime256v1 -genkey
//
// Regression test for https://github.com/hashicorp/vault/issues/16667
resp, err := CBWrite(b, s, "config/ca", map[string]interface{}{
"pem_bundle": `
-----BEGIN EC PARAMETERS-----
BggqhkjOPQMBBw==
-----END EC PARAMETERS-----
-----BEGIN EC PRIVATE KEY-----
MHcCAQEEINzXthCZdhyV7+wIEBl/ty+ctNsUS99ykTeax6EbYZtvoAoGCCqGSM49
AwEHoUQDQgAE57NX8bR/nDoW8yRgLswoXBQcjHrdyfuHS0gPwki6BNnfunUzryVb
8f22/JWj6fsEF6AOADZlrswKIbR2Es9e/w==
-----END EC PRIVATE KEY-----
`,
})
require.NoError(t, err)
require.NotNil(t, resp, "expected ca info")
importedKeys := resp.Data["imported_keys"].([]string)
importedIssuers := resp.Data["imported_issuers"].([]string)
require.Equal(t, len(importedKeys), 1)
require.Equal(t, len(importedIssuers), 0)
}
Add per-issuer AIA URI information to PKI secrets engine (#16563) * Add per-issuer AIA URI information Per discussion on GitHub with @maxb, this allows issuers to have their own copy of AIA URIs. Because each issuer has its own URLs (for CA and CRL access), its necessary to mint their issued certs pointing to the correct issuer and not to the global default issuer. For anyone using multiple issuers within a mount, this change allows the issuer to point back to itself via leaf's AIA info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation on per-issuer AIA info Also add it to the considerations page as something to watch out for. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for per-issuer AIA information Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor AIA setting on the issuer This introduces a common helper per Steve's suggestion. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error messages w.r.t. AIA naming Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error messages regarding AIA URLs This clarifies which request parameter the invalid URL is contained in, disambiguating the sometimes ambiguous usage of AIA, per suggestion by Max. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename getURLs -> getGlobalAIAURLs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Correct AIA acronym expansion word orders Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix bad comment suggesting re-generating roots Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add two entries to URL tests Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-19 15:43:44 +00:00
func TestPerIssuerAIA(t *testing.T) {
t.Parallel()
b, s := CreateBackendWithStorage(t)
Add per-issuer AIA URI information to PKI secrets engine (#16563) * Add per-issuer AIA URI information Per discussion on GitHub with @maxb, this allows issuers to have their own copy of AIA URIs. Because each issuer has its own URLs (for CA and CRL access), its necessary to mint their issued certs pointing to the correct issuer and not to the global default issuer. For anyone using multiple issuers within a mount, this change allows the issuer to point back to itself via leaf's AIA info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation on per-issuer AIA info Also add it to the considerations page as something to watch out for. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for per-issuer AIA information Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor AIA setting on the issuer This introduces a common helper per Steve's suggestion. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error messages w.r.t. AIA naming Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error messages regarding AIA URLs This clarifies which request parameter the invalid URL is contained in, disambiguating the sometimes ambiguous usage of AIA, per suggestion by Max. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename getURLs -> getGlobalAIAURLs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Correct AIA acronym expansion word orders Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix bad comment suggesting re-generating roots Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add two entries to URL tests Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-19 15:43:44 +00:00
// Generating a root without anything should not have AIAs.
resp, err := CBWrite(b, s, "root/generate/internal", map[string]interface{}{
"common_name": "root example.com",
"issuer_name": "root",
"key_type": "ec",
})
require.NoError(t, err)
require.NotNil(t, resp)
rootCert := parseCert(t, resp.Data["certificate"].(string))
require.Empty(t, rootCert.OCSPServer)
require.Empty(t, rootCert.IssuingCertificateURL)
require.Empty(t, rootCert.CRLDistributionPoints)
// Set some local URLs on the issuer.
_, err = CBWrite(b, s, "issuer/default", map[string]interface{}{
"issuing_certificates": []string{"https://google.com"},
})
require.NoError(t, err)
_, err = CBWrite(b, s, "roles/testing", map[string]interface{}{
"allow_any_name": true,
"ttl": "85s",
"key_type": "ec",
})
require.NoError(t, err)
// Issue something with this re-configured issuer.
resp, err = CBWrite(b, s, "issuer/default/issue/testing", map[string]interface{}{
"common_name": "localhost.com",
})
require.NoError(t, err)
require.NotNil(t, resp)
leafCert := parseCert(t, resp.Data["certificate"].(string))
require.Empty(t, leafCert.OCSPServer)
require.Equal(t, leafCert.IssuingCertificateURL, []string{"https://google.com"})
require.Empty(t, leafCert.CRLDistributionPoints)
// Set global URLs and ensure they don't appear on this issuer's leaf.
_, err = CBWrite(b, s, "config/urls", map[string]interface{}{
"issuing_certificates": []string{"https://example.com/ca", "https://backup.example.com/ca"},
"crl_distribution_points": []string{"https://example.com/crl", "https://backup.example.com/crl"},
"ocsp_servers": []string{"https://example.com/ocsp", "https://backup.example.com/ocsp"},
})
require.NoError(t, err)
resp, err = CBWrite(b, s, "issuer/default/issue/testing", map[string]interface{}{
"common_name": "localhost.com",
})
require.NoError(t, err)
require.NotNil(t, resp)
leafCert = parseCert(t, resp.Data["certificate"].(string))
require.Empty(t, leafCert.OCSPServer)
require.Equal(t, leafCert.IssuingCertificateURL, []string{"https://google.com"})
require.Empty(t, leafCert.CRLDistributionPoints)
// Now come back and remove the local modifications and ensure we get
// the defaults again.
_, err = CBPatch(b, s, "issuer/default", map[string]interface{}{
"issuing_certificates": []string{},
})
require.NoError(t, err)
resp, err = CBWrite(b, s, "issuer/default/issue/testing", map[string]interface{}{
"common_name": "localhost.com",
})
require.NoError(t, err)
require.NotNil(t, resp)
leafCert = parseCert(t, resp.Data["certificate"].(string))
require.Equal(t, leafCert.IssuingCertificateURL, []string{"https://example.com/ca", "https://backup.example.com/ca"})
require.Equal(t, leafCert.OCSPServer, []string{"https://example.com/ocsp", "https://backup.example.com/ocsp"})
require.Equal(t, leafCert.CRLDistributionPoints, []string{"https://example.com/crl", "https://backup.example.com/crl"})
// Validate that we can set an issuer name and remove it.
_, err = CBPatch(b, s, "issuer/default", map[string]interface{}{
"issuer_name": "my-issuer",
})
require.NoError(t, err)
_, err = CBPatch(b, s, "issuer/default", map[string]interface{}{
"issuer_name": "",
})
require.NoError(t, err)
Add per-issuer AIA URI information to PKI secrets engine (#16563) * Add per-issuer AIA URI information Per discussion on GitHub with @maxb, this allows issuers to have their own copy of AIA URIs. Because each issuer has its own URLs (for CA and CRL access), its necessary to mint their issued certs pointing to the correct issuer and not to the global default issuer. For anyone using multiple issuers within a mount, this change allows the issuer to point back to itself via leaf's AIA info. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation on per-issuer AIA info Also add it to the considerations page as something to watch out for. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for per-issuer AIA information Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor AIA setting on the issuer This introduces a common helper per Steve's suggestion. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error messages w.r.t. AIA naming Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Clarify error messages regarding AIA URLs This clarifies which request parameter the invalid URL is contained in, disambiguating the sometimes ambiguous usage of AIA, per suggestion by Max. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rename getURLs -> getGlobalAIAURLs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Correct AIA acronym expansion word orders Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Fix bad comment suggesting re-generating roots Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add two entries to URL tests Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-19 15:43:44 +00:00
}
func TestIssuersWithoutCRLBits(t *testing.T) {
t.Parallel()
b, s := CreateBackendWithStorage(t)
// Importing a root without CRL signing bits should work fine.
customBundleWithoutCRLBits := `
-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----
-----BEGIN PRIVATE KEY-----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-----END PRIVATE KEY-----
`
resp, err := CBWrite(b, s, "issuers/import/bundle", map[string]interface{}{
"pem_bundle": customBundleWithoutCRLBits,
})
require.NoError(t, err)
require.NotNil(t, resp)
require.NotEmpty(t, resp.Data)
require.NotEmpty(t, resp.Data["imported_issuers"])
require.NotEmpty(t, resp.Data["imported_keys"])
require.NotEmpty(t, resp.Data["mapping"])
// Shouldn't have crl-signing on the newly imported issuer's usage.
resp, err = CBRead(b, s, "issuer/default")
require.NoError(t, err)
require.NotNil(t, resp)
require.NotEmpty(t, resp.Data)
require.NotEmpty(t, resp.Data["usage"])
require.NotContains(t, resp.Data["usage"], "crl-signing")
// Modifying to set CRL should fail.
resp, err = CBPatch(b, s, "issuer/default", map[string]interface{}{
"usage": "issuing-certificates,crl-signing",
})
require.Error(t, err)
require.True(t, resp.IsError())
// Modifying to set issuing-certificates and ocsp-signing should succeed.
resp, err = CBPatch(b, s, "issuer/default", map[string]interface{}{
"usage": "issuing-certificates,ocsp-signing",
})
require.NoError(t, err)
require.NotNil(t, resp)
require.NotEmpty(t, resp.Data)
require.NotEmpty(t, resp.Data["usage"])
require.NotContains(t, resp.Data["usage"], "crl-signing")
}
PKI - Honor header If-Modified-Since if present (#16249) * honor header if-modified-since if present * pathGetIssuerCRL first version * check if modified since for CA endpoints * fix date comparison for CA endpoints * suggested changes and refactoring * add writeIssuer to updateDefaultIssuerId and fix error * Move methods out of storage.go into util.go For the most part, these take a SC as param, but aren't directly storage relevant operations. Move them out of storage.go as a result. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use UTC timezone for storage Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rework path_fetch for better if-modified-since handling Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Invalidate all issuers, CRLs on default write When the default is updated, access under earlier timestamps will not work as we're unclear if the timestamp is for this issuer or a previous issuer. Thus, we need to invalidate the CRL and both issuers involved (previous, next) by updating their LastModifiedTimes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for If-Modified-Since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Correctly invalidate default issuer changes When the default issuer changes, we'll have to mark the invalidation on PR secondary clusters, so they know to update their CRL mapping as well. The swapped issuers will have an updated modification time (which will eventually replicate down and thus be correct), but the CRL modification time is cluster-local information and thus won't be replicated. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * make fmt Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sendNotModifiedResponseIfNecessary Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation on if-modified-since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-29 19:28:47 +00:00
func TestBackend_IfModifiedSinceHeaders(t *testing.T) {
t.Parallel()
coreConfig := &vault.CoreConfig{
LogicalBackends: map[string]logical.Factory{
"pki": Factory,
},
}
cluster := vault.NewTestCluster(t, coreConfig, &vault.TestClusterOptions{
HandlerFunc: vaulthttp.Handler,
})
cluster.Start()
defer cluster.Cleanup()
client := cluster.Cores[0].Client
// Mount PKI.
err := client.Sys().Mount("pki", &api.MountInput{
Type: "pki",
Config: api.MountConfigInput{
DefaultLeaseTTL: "16h",
MaxLeaseTTL: "60h",
// Required to allow the header to be passed through.
PassthroughRequestHeaders: []string{"if-modified-since"},
AllowedResponseHeaders: []string{"Last-Modified"},
PKI - Honor header If-Modified-Since if present (#16249) * honor header if-modified-since if present * pathGetIssuerCRL first version * check if modified since for CA endpoints * fix date comparison for CA endpoints * suggested changes and refactoring * add writeIssuer to updateDefaultIssuerId and fix error * Move methods out of storage.go into util.go For the most part, these take a SC as param, but aren't directly storage relevant operations. Move them out of storage.go as a result. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use UTC timezone for storage Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rework path_fetch for better if-modified-since handling Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Invalidate all issuers, CRLs on default write When the default is updated, access under earlier timestamps will not work as we're unclear if the timestamp is for this issuer or a previous issuer. Thus, we need to invalidate the CRL and both issuers involved (previous, next) by updating their LastModifiedTimes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for If-Modified-Since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Correctly invalidate default issuer changes When the default issuer changes, we'll have to mark the invalidation on PR secondary clusters, so they know to update their CRL mapping as well. The swapped issuers will have an updated modification time (which will eventually replicate down and thus be correct), but the CRL modification time is cluster-local information and thus won't be replicated. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * make fmt Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sendNotModifiedResponseIfNecessary Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation on if-modified-since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-29 19:28:47 +00:00
},
})
require.NoError(t, err)
// Get a time before CA generation. Subtract two seconds to ensure
// the value in the seconds field is different than the time the CA
// is actually generated at.
beforeOldCAGeneration := time.Now().Add(-2 * time.Second)
// Generate an internal CA. This one is the default.
resp, err := client.Logical().Write("pki/root/generate/internal", map[string]interface{}{
"ttl": "40h",
"common_name": "Root X1",
"key_type": "ec",
"issuer_name": "old-root",
})
require.NoError(t, err)
require.NotNil(t, resp)
require.NotNil(t, resp.Data)
require.NotEmpty(t, resp.Data["certificate"])
// CA is generated, but give a grace window.
afterOldCAGeneration := time.Now().Add(2 * time.Second)
// When you _save_ headers, client returns a copy. But when you go to
// reset them, it doesn't create a new copy (and instead directly
// assigns). This means we have to continually refresh our view of the
// last headers, otherwise the headers added after the last set operation
// leak into this copy... Yuck!
lastHeaders := client.Headers()
for _, path := range []string{"pki/cert/ca", "pki/cert/crl", "pki/issuer/default/json", "pki/issuer/old-root/json", "pki/issuer/old-root/crl", "pki/cert/delta-crl", "pki/issuer/old-root/crl/delta"} {
PKI - Honor header If-Modified-Since if present (#16249) * honor header if-modified-since if present * pathGetIssuerCRL first version * check if modified since for CA endpoints * fix date comparison for CA endpoints * suggested changes and refactoring * add writeIssuer to updateDefaultIssuerId and fix error * Move methods out of storage.go into util.go For the most part, these take a SC as param, but aren't directly storage relevant operations. Move them out of storage.go as a result. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use UTC timezone for storage Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rework path_fetch for better if-modified-since handling Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Invalidate all issuers, CRLs on default write When the default is updated, access under earlier timestamps will not work as we're unclear if the timestamp is for this issuer or a previous issuer. Thus, we need to invalidate the CRL and both issuers involved (previous, next) by updating their LastModifiedTimes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for If-Modified-Since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Correctly invalidate default issuer changes When the default issuer changes, we'll have to mark the invalidation on PR secondary clusters, so they know to update their CRL mapping as well. The swapped issuers will have an updated modification time (which will eventually replicate down and thus be correct), but the CRL modification time is cluster-local information and thus won't be replicated. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * make fmt Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sendNotModifiedResponseIfNecessary Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation on if-modified-since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-29 19:28:47 +00:00
t.Logf("path: %v", path)
field := "certificate"
if strings.HasPrefix(path, "pki/issuer") && strings.Contains(path, "/crl") {
PKI - Honor header If-Modified-Since if present (#16249) * honor header if-modified-since if present * pathGetIssuerCRL first version * check if modified since for CA endpoints * fix date comparison for CA endpoints * suggested changes and refactoring * add writeIssuer to updateDefaultIssuerId and fix error * Move methods out of storage.go into util.go For the most part, these take a SC as param, but aren't directly storage relevant operations. Move them out of storage.go as a result. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use UTC timezone for storage Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rework path_fetch for better if-modified-since handling Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Invalidate all issuers, CRLs on default write When the default is updated, access under earlier timestamps will not work as we're unclear if the timestamp is for this issuer or a previous issuer. Thus, we need to invalidate the CRL and both issuers involved (previous, next) by updating their LastModifiedTimes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for If-Modified-Since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Correctly invalidate default issuer changes When the default issuer changes, we'll have to mark the invalidation on PR secondary clusters, so they know to update their CRL mapping as well. The swapped issuers will have an updated modification time (which will eventually replicate down and thus be correct), but the CRL modification time is cluster-local information and thus won't be replicated. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * make fmt Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sendNotModifiedResponseIfNecessary Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation on if-modified-since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-29 19:28:47 +00:00
field = "crl"
}
// Reading the CA should work, without a header.
resp, err := client.Logical().Read(path)
require.NoError(t, err)
require.NotNil(t, resp)
require.NotNil(t, resp.Data)
require.NotEmpty(t, resp.Data[field])
// Ensure that the CA is returned correctly if we give it the old time.
client.AddHeader("If-Modified-Since", beforeOldCAGeneration.Format(time.RFC1123))
resp, err = client.Logical().Read(path)
require.NoError(t, err)
require.NotNil(t, resp)
require.NotNil(t, resp.Data)
require.NotEmpty(t, resp.Data[field])
client.SetHeaders(lastHeaders)
lastHeaders = client.Headers()
// Ensure that the CA is elided if we give it the present time (plus a
// grace window).
client.AddHeader("If-Modified-Since", afterOldCAGeneration.Format(time.RFC1123))
t.Logf("headers: %v", client.Headers())
resp, err = client.Logical().Read(path)
require.NoError(t, err)
require.Nil(t, resp)
client.SetHeaders(lastHeaders)
lastHeaders = client.Headers()
}
// Wait three seconds. This ensures we have adequate grace period
// to distinguish the two cases, even with grace periods.
time.Sleep(3 * time.Second)
// Generating a second root. This one isn't the default.
beforeNewCAGeneration := time.Now().Add(-2 * time.Second)
// Generate an internal CA. This one is the default.
_, err = client.Logical().Write("pki/root/generate/internal", map[string]interface{}{
"ttl": "40h",
"common_name": "Root X1",
"key_type": "ec",
"issuer_name": "new-root",
})
require.NoError(t, err)
// As above.
afterNewCAGeneration := time.Now().Add(2 * time.Second)
// New root isn't the default, so it has fewer paths.
for _, path := range []string{"pki/issuer/new-root/json", "pki/issuer/new-root/crl", "pki/issuer/new-root/crl/delta"} {
PKI - Honor header If-Modified-Since if present (#16249) * honor header if-modified-since if present * pathGetIssuerCRL first version * check if modified since for CA endpoints * fix date comparison for CA endpoints * suggested changes and refactoring * add writeIssuer to updateDefaultIssuerId and fix error * Move methods out of storage.go into util.go For the most part, these take a SC as param, but aren't directly storage relevant operations. Move them out of storage.go as a result. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use UTC timezone for storage Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rework path_fetch for better if-modified-since handling Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Invalidate all issuers, CRLs on default write When the default is updated, access under earlier timestamps will not work as we're unclear if the timestamp is for this issuer or a previous issuer. Thus, we need to invalidate the CRL and both issuers involved (previous, next) by updating their LastModifiedTimes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for If-Modified-Since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Correctly invalidate default issuer changes When the default issuer changes, we'll have to mark the invalidation on PR secondary clusters, so they know to update their CRL mapping as well. The swapped issuers will have an updated modification time (which will eventually replicate down and thus be correct), but the CRL modification time is cluster-local information and thus won't be replicated. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * make fmt Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sendNotModifiedResponseIfNecessary Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation on if-modified-since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-29 19:28:47 +00:00
t.Logf("path: %v", path)
field := "certificate"
if strings.HasPrefix(path, "pki/issuer") && strings.Contains(path, "/crl") {
PKI - Honor header If-Modified-Since if present (#16249) * honor header if-modified-since if present * pathGetIssuerCRL first version * check if modified since for CA endpoints * fix date comparison for CA endpoints * suggested changes and refactoring * add writeIssuer to updateDefaultIssuerId and fix error * Move methods out of storage.go into util.go For the most part, these take a SC as param, but aren't directly storage relevant operations. Move them out of storage.go as a result. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use UTC timezone for storage Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rework path_fetch for better if-modified-since handling Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Invalidate all issuers, CRLs on default write When the default is updated, access under earlier timestamps will not work as we're unclear if the timestamp is for this issuer or a previous issuer. Thus, we need to invalidate the CRL and both issuers involved (previous, next) by updating their LastModifiedTimes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for If-Modified-Since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Correctly invalidate default issuer changes When the default issuer changes, we'll have to mark the invalidation on PR secondary clusters, so they know to update their CRL mapping as well. The swapped issuers will have an updated modification time (which will eventually replicate down and thus be correct), but the CRL modification time is cluster-local information and thus won't be replicated. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * make fmt Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sendNotModifiedResponseIfNecessary Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation on if-modified-since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-29 19:28:47 +00:00
field = "crl"
}
// Reading the CA should work, without a header.
resp, err := client.Logical().Read(path)
require.NoError(t, err)
require.NotNil(t, resp)
require.NotNil(t, resp.Data)
require.NotEmpty(t, resp.Data[field])
// Ensure that the CA is returned correctly if we give it the old time.
client.AddHeader("If-Modified-Since", beforeNewCAGeneration.Format(time.RFC1123))
resp, err = client.Logical().Read(path)
require.NoError(t, err)
require.NotNil(t, resp)
require.NotNil(t, resp.Data)
require.NotEmpty(t, resp.Data[field])
client.SetHeaders(lastHeaders)
lastHeaders = client.Headers()
// Ensure that the CA is elided if we give it the present time (plus a
// grace window).
client.AddHeader("If-Modified-Since", afterNewCAGeneration.Format(time.RFC1123))
t.Logf("headers: %v", client.Headers())
resp, err = client.Logical().Read(path)
require.NoError(t, err)
require.Nil(t, resp)
client.SetHeaders(lastHeaders)
lastHeaders = client.Headers()
}
// Wait three seconds. This ensures we have adequate grace period
// to distinguish the two cases, even with grace periods.
time.Sleep(3 * time.Second)
// Now swap the default issuers around.
_, err = client.Logical().Write("pki/config/issuers", map[string]interface{}{
"default": "new-root",
})
require.NoError(t, err)
// Reading both with the last modified date should return new values.
for _, path := range []string{"pki/cert/ca", "pki/cert/crl", "pki/issuer/default/json", "pki/issuer/old-root/json", "pki/issuer/new-root/json", "pki/issuer/old-root/crl", "pki/issuer/new-root/crl", "pki/cert/delta-crl", "pki/issuer/old-root/crl/delta", "pki/issuer/new-root/crl/delta"} {
PKI - Honor header If-Modified-Since if present (#16249) * honor header if-modified-since if present * pathGetIssuerCRL first version * check if modified since for CA endpoints * fix date comparison for CA endpoints * suggested changes and refactoring * add writeIssuer to updateDefaultIssuerId and fix error * Move methods out of storage.go into util.go For the most part, these take a SC as param, but aren't directly storage relevant operations. Move them out of storage.go as a result. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use UTC timezone for storage Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rework path_fetch for better if-modified-since handling Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Invalidate all issuers, CRLs on default write When the default is updated, access under earlier timestamps will not work as we're unclear if the timestamp is for this issuer or a previous issuer. Thus, we need to invalidate the CRL and both issuers involved (previous, next) by updating their LastModifiedTimes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for If-Modified-Since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Correctly invalidate default issuer changes When the default issuer changes, we'll have to mark the invalidation on PR secondary clusters, so they know to update their CRL mapping as well. The swapped issuers will have an updated modification time (which will eventually replicate down and thus be correct), but the CRL modification time is cluster-local information and thus won't be replicated. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * make fmt Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sendNotModifiedResponseIfNecessary Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation on if-modified-since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-29 19:28:47 +00:00
t.Logf("path: %v", path)
field := "certificate"
if strings.HasPrefix(path, "pki/issuer") && strings.Contains(path, "/crl") {
PKI - Honor header If-Modified-Since if present (#16249) * honor header if-modified-since if present * pathGetIssuerCRL first version * check if modified since for CA endpoints * fix date comparison for CA endpoints * suggested changes and refactoring * add writeIssuer to updateDefaultIssuerId and fix error * Move methods out of storage.go into util.go For the most part, these take a SC as param, but aren't directly storage relevant operations. Move them out of storage.go as a result. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use UTC timezone for storage Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rework path_fetch for better if-modified-since handling Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Invalidate all issuers, CRLs on default write When the default is updated, access under earlier timestamps will not work as we're unclear if the timestamp is for this issuer or a previous issuer. Thus, we need to invalidate the CRL and both issuers involved (previous, next) by updating their LastModifiedTimes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for If-Modified-Since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Correctly invalidate default issuer changes When the default issuer changes, we'll have to mark the invalidation on PR secondary clusters, so they know to update their CRL mapping as well. The swapped issuers will have an updated modification time (which will eventually replicate down and thus be correct), but the CRL modification time is cluster-local information and thus won't be replicated. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * make fmt Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sendNotModifiedResponseIfNecessary Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation on if-modified-since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-29 19:28:47 +00:00
field = "crl"
}
// Ensure that the CA is returned correctly if we give it the old time.
client.AddHeader("If-Modified-Since", afterOldCAGeneration.Format(time.RFC1123))
resp, err = client.Logical().Read(path)
require.NoError(t, err)
require.NotNil(t, resp)
require.NotNil(t, resp.Data)
require.NotEmpty(t, resp.Data[field])
client.SetHeaders(lastHeaders)
lastHeaders = client.Headers()
// Ensure that the CA is returned correctly if we give it the old time.
client.AddHeader("If-Modified-Since", afterNewCAGeneration.Format(time.RFC1123))
resp, err = client.Logical().Read(path)
require.NoError(t, err)
require.NotNil(t, resp)
require.NotNil(t, resp.Data)
require.NotEmpty(t, resp.Data[field])
client.SetHeaders(lastHeaders)
lastHeaders = client.Headers()
}
// Wait for things to settle, record the present time, and wait for the
// clock to definitely tick over again.
time.Sleep(2 * time.Second)
preRevocationTimestamp := time.Now()
time.Sleep(2 * time.Second)
// The above tests should say everything is cached.
for _, path := range []string{"pki/cert/ca", "pki/cert/crl", "pki/issuer/default/json", "pki/issuer/old-root/json", "pki/issuer/new-root/json", "pki/issuer/old-root/crl", "pki/issuer/new-root/crl", "pki/cert/delta-crl", "pki/issuer/old-root/crl/delta", "pki/issuer/new-root/crl/delta"} {
PKI - Honor header If-Modified-Since if present (#16249) * honor header if-modified-since if present * pathGetIssuerCRL first version * check if modified since for CA endpoints * fix date comparison for CA endpoints * suggested changes and refactoring * add writeIssuer to updateDefaultIssuerId and fix error * Move methods out of storage.go into util.go For the most part, these take a SC as param, but aren't directly storage relevant operations. Move them out of storage.go as a result. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use UTC timezone for storage Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rework path_fetch for better if-modified-since handling Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Invalidate all issuers, CRLs on default write When the default is updated, access under earlier timestamps will not work as we're unclear if the timestamp is for this issuer or a previous issuer. Thus, we need to invalidate the CRL and both issuers involved (previous, next) by updating their LastModifiedTimes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for If-Modified-Since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Correctly invalidate default issuer changes When the default issuer changes, we'll have to mark the invalidation on PR secondary clusters, so they know to update their CRL mapping as well. The swapped issuers will have an updated modification time (which will eventually replicate down and thus be correct), but the CRL modification time is cluster-local information and thus won't be replicated. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * make fmt Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sendNotModifiedResponseIfNecessary Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation on if-modified-since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-29 19:28:47 +00:00
t.Logf("path: %v", path)
// Ensure that the CA is returned correctly if we give it the new time.
client.AddHeader("If-Modified-Since", preRevocationTimestamp.Format(time.RFC1123))
resp, err = client.Logical().Read(path)
require.NoError(t, err)
require.Nil(t, resp)
client.SetHeaders(lastHeaders)
lastHeaders = client.Headers()
}
// We could generate some leaves and verify the revocation updates the
// CRL. But, revoking the issuer behaves the same, so let's do that
// instead.
_, err = client.Logical().Write("pki/issuer/old-root/revoke", map[string]interface{}{})
require.NoError(t, err)
// CA should still be valid.
for _, path := range []string{"pki/cert/ca", "pki/issuer/default/json", "pki/issuer/old-root/json", "pki/issuer/new-root/json"} {
t.Logf("path: %v", path)
// Ensure that the CA is returned correctly if we give it the old time.
client.AddHeader("If-Modified-Since", preRevocationTimestamp.Format(time.RFC1123))
resp, err = client.Logical().Read(path)
require.NoError(t, err)
require.Nil(t, resp)
client.SetHeaders(lastHeaders)
lastHeaders = client.Headers()
}
// CRL should be invalidated
for _, path := range []string{"pki/cert/crl", "pki/issuer/old-root/crl", "pki/issuer/new-root/crl", "pki/cert/delta-crl", "pki/issuer/old-root/crl/delta", "pki/issuer/new-root/crl/delta"} {
PKI - Honor header If-Modified-Since if present (#16249) * honor header if-modified-since if present * pathGetIssuerCRL first version * check if modified since for CA endpoints * fix date comparison for CA endpoints * suggested changes and refactoring * add writeIssuer to updateDefaultIssuerId and fix error * Move methods out of storage.go into util.go For the most part, these take a SC as param, but aren't directly storage relevant operations. Move them out of storage.go as a result. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use UTC timezone for storage Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rework path_fetch for better if-modified-since handling Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Invalidate all issuers, CRLs on default write When the default is updated, access under earlier timestamps will not work as we're unclear if the timestamp is for this issuer or a previous issuer. Thus, we need to invalidate the CRL and both issuers involved (previous, next) by updating their LastModifiedTimes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for If-Modified-Since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Correctly invalidate default issuer changes When the default issuer changes, we'll have to mark the invalidation on PR secondary clusters, so they know to update their CRL mapping as well. The swapped issuers will have an updated modification time (which will eventually replicate down and thus be correct), but the CRL modification time is cluster-local information and thus won't be replicated. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * make fmt Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sendNotModifiedResponseIfNecessary Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation on if-modified-since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-29 19:28:47 +00:00
t.Logf("path: %v", path)
field := "certificate"
if strings.HasPrefix(path, "pki/issuer") && strings.Contains(path, "/crl") {
PKI - Honor header If-Modified-Since if present (#16249) * honor header if-modified-since if present * pathGetIssuerCRL first version * check if modified since for CA endpoints * fix date comparison for CA endpoints * suggested changes and refactoring * add writeIssuer to updateDefaultIssuerId and fix error * Move methods out of storage.go into util.go For the most part, these take a SC as param, but aren't directly storage relevant operations. Move them out of storage.go as a result. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use UTC timezone for storage Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rework path_fetch for better if-modified-since handling Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Invalidate all issuers, CRLs on default write When the default is updated, access under earlier timestamps will not work as we're unclear if the timestamp is for this issuer or a previous issuer. Thus, we need to invalidate the CRL and both issuers involved (previous, next) by updating their LastModifiedTimes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for If-Modified-Since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Correctly invalidate default issuer changes When the default issuer changes, we'll have to mark the invalidation on PR secondary clusters, so they know to update their CRL mapping as well. The swapped issuers will have an updated modification time (which will eventually replicate down and thus be correct), but the CRL modification time is cluster-local information and thus won't be replicated. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * make fmt Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sendNotModifiedResponseIfNecessary Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation on if-modified-since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-29 19:28:47 +00:00
field = "crl"
}
client.AddHeader("If-Modified-Since", preRevocationTimestamp.Format(time.RFC1123))
resp, err = client.Logical().Read(path)
require.NoError(t, err)
require.NotNil(t, resp)
require.NotNil(t, resp.Data)
require.NotEmpty(t, resp.Data[field])
client.SetHeaders(lastHeaders)
lastHeaders = client.Headers()
}
// If we send some time in the future, everything should be cached again!
PKI - Honor header If-Modified-Since if present (#16249) * honor header if-modified-since if present * pathGetIssuerCRL first version * check if modified since for CA endpoints * fix date comparison for CA endpoints * suggested changes and refactoring * add writeIssuer to updateDefaultIssuerId and fix error * Move methods out of storage.go into util.go For the most part, these take a SC as param, but aren't directly storage relevant operations. Move them out of storage.go as a result. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use UTC timezone for storage Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rework path_fetch for better if-modified-since handling Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Invalidate all issuers, CRLs on default write When the default is updated, access under earlier timestamps will not work as we're unclear if the timestamp is for this issuer or a previous issuer. Thus, we need to invalidate the CRL and both issuers involved (previous, next) by updating their LastModifiedTimes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for If-Modified-Since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Correctly invalidate default issuer changes When the default issuer changes, we'll have to mark the invalidation on PR secondary clusters, so they know to update their CRL mapping as well. The swapped issuers will have an updated modification time (which will eventually replicate down and thus be correct), but the CRL modification time is cluster-local information and thus won't be replicated. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * make fmt Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sendNotModifiedResponseIfNecessary Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation on if-modified-since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-29 19:28:47 +00:00
futureTime := time.Now().Add(30 * time.Second)
for _, path := range []string{"pki/cert/ca", "pki/cert/crl", "pki/issuer/default/json", "pki/issuer/old-root/json", "pki/issuer/new-root/json", "pki/issuer/old-root/crl", "pki/issuer/new-root/crl", "pki/cert/delta-crl", "pki/issuer/old-root/crl/delta", "pki/issuer/new-root/crl/delta"} {
PKI - Honor header If-Modified-Since if present (#16249) * honor header if-modified-since if present * pathGetIssuerCRL first version * check if modified since for CA endpoints * fix date comparison for CA endpoints * suggested changes and refactoring * add writeIssuer to updateDefaultIssuerId and fix error * Move methods out of storage.go into util.go For the most part, these take a SC as param, but aren't directly storage relevant operations. Move them out of storage.go as a result. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use UTC timezone for storage Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rework path_fetch for better if-modified-since handling Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Invalidate all issuers, CRLs on default write When the default is updated, access under earlier timestamps will not work as we're unclear if the timestamp is for this issuer or a previous issuer. Thus, we need to invalidate the CRL and both issuers involved (previous, next) by updating their LastModifiedTimes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for If-Modified-Since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Correctly invalidate default issuer changes When the default issuer changes, we'll have to mark the invalidation on PR secondary clusters, so they know to update their CRL mapping as well. The swapped issuers will have an updated modification time (which will eventually replicate down and thus be correct), but the CRL modification time is cluster-local information and thus won't be replicated. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * make fmt Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sendNotModifiedResponseIfNecessary Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation on if-modified-since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-29 19:28:47 +00:00
t.Logf("path: %v", path)
// Ensure that the CA is returned correctly if we give it the new time.
client.AddHeader("If-Modified-Since", futureTime.Format(time.RFC1123))
resp, err = client.Logical().Read(path)
require.NoError(t, err)
require.Nil(t, resp)
client.SetHeaders(lastHeaders)
lastHeaders = client.Headers()
}
beforeThreeWaySwap := time.Now().Add(-2 * time.Second)
// Now, do a three-way swap of names (old->tmp; new->old; tmp->new). This
// should result in all names/CRLs being invalidated.
_, err = client.Logical().JSONMergePatch(ctx, "pki/issuer/old-root", map[string]interface{}{
"issuer_name": "tmp-root",
})
require.NoError(t, err)
_, err = client.Logical().JSONMergePatch(ctx, "pki/issuer/new-root", map[string]interface{}{
"issuer_name": "old-root",
})
require.NoError(t, err)
_, err = client.Logical().JSONMergePatch(ctx, "pki/issuer/tmp-root", map[string]interface{}{
"issuer_name": "new-root",
})
require.NoError(t, err)
afterThreeWaySwap := time.Now().Add(2 * time.Second)
for _, path := range []string{"pki/cert/ca", "pki/cert/crl", "pki/issuer/default/json", "pki/issuer/old-root/json", "pki/issuer/new-root/json", "pki/issuer/old-root/crl", "pki/issuer/new-root/crl", "pki/cert/delta-crl", "pki/issuer/old-root/crl/delta", "pki/issuer/new-root/crl/delta"} {
t.Logf("path: %v", path)
field := "certificate"
if strings.HasPrefix(path, "pki/issuer") && strings.Contains(path, "/crl") {
field = "crl"
}
// Ensure that the CA is returned if we give it the pre-update time.
client.AddHeader("If-Modified-Since", beforeThreeWaySwap.Format(time.RFC1123))
resp, err = client.Logical().Read(path)
require.NoError(t, err)
require.NotNil(t, resp)
require.NotNil(t, resp.Data)
require.NotEmpty(t, resp.Data[field])
client.SetHeaders(lastHeaders)
lastHeaders = client.Headers()
// Ensure that the CA is elided correctly if we give it the after time.
client.AddHeader("If-Modified-Since", afterThreeWaySwap.Format(time.RFC1123))
resp, err = client.Logical().Read(path)
require.NoError(t, err)
require.Nil(t, resp)
client.SetHeaders(lastHeaders)
lastHeaders = client.Headers()
}
// Finally, rebuild the delta CRL and ensure that only that is
// invalidated. We first need to enable it though, and wait for
// all CRLs to rebuild.
_, err = client.Logical().Write("pki/config/crl", map[string]interface{}{
"auto_rebuild": true,
"enable_delta": true,
})
require.NoError(t, err)
time.Sleep(4 * time.Second)
beforeDeltaRotation := time.Now().Add(-2 * time.Second)
resp, err = client.Logical().Read("pki/crl/rotate-delta")
require.NoError(t, err)
require.NotNil(t, resp)
require.NotNil(t, resp.Data)
require.Equal(t, resp.Data["success"], true)
afterDeltaRotation := time.Now().Add(2 * time.Second)
for _, path := range []string{"pki/cert/ca", "pki/cert/crl", "pki/issuer/default/json", "pki/issuer/old-root/json", "pki/issuer/new-root/json", "pki/issuer/old-root/crl", "pki/issuer/new-root/crl"} {
t.Logf("path: %v", path)
for _, when := range []time.Time{beforeDeltaRotation, afterDeltaRotation} {
client.AddHeader("If-Modified-Since", when.Format(time.RFC1123))
resp, err = client.Logical().Read(path)
require.NoError(t, err)
require.Nil(t, resp)
client.SetHeaders(lastHeaders)
lastHeaders = client.Headers()
}
}
for _, path := range []string{"pki/cert/delta-crl", "pki/issuer/old-root/crl/delta", "pki/issuer/new-root/crl/delta"} {
t.Logf("path: %v", path)
field := "certificate"
if strings.HasPrefix(path, "pki/issuer") && strings.Contains(path, "/crl") {
field = "crl"
}
// Ensure that the CRL is present if we give it the pre-update time.
client.AddHeader("If-Modified-Since", beforeDeltaRotation.Format(time.RFC1123))
resp, err = client.Logical().Read(path)
require.NoError(t, err)
require.NotNil(t, resp)
require.NotNil(t, resp.Data)
require.NotEmpty(t, resp.Data[field])
client.SetHeaders(lastHeaders)
lastHeaders = client.Headers()
client.AddHeader("If-Modified-Since", afterDeltaRotation.Format(time.RFC1123))
resp, err = client.Logical().Read(path)
require.NoError(t, err)
require.Nil(t, resp)
client.SetHeaders(lastHeaders)
lastHeaders = client.Headers()
}
PKI - Honor header If-Modified-Since if present (#16249) * honor header if-modified-since if present * pathGetIssuerCRL first version * check if modified since for CA endpoints * fix date comparison for CA endpoints * suggested changes and refactoring * add writeIssuer to updateDefaultIssuerId and fix error * Move methods out of storage.go into util.go For the most part, these take a SC as param, but aren't directly storage relevant operations. Move them out of storage.go as a result. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Use UTC timezone for storage Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Rework path_fetch for better if-modified-since handling Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Invalidate all issuers, CRLs on default write When the default is updated, access under earlier timestamps will not work as we're unclear if the timestamp is for this issuer or a previous issuer. Thus, we need to invalidate the CRL and both issuers involved (previous, next) by updating their LastModifiedTimes. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests for If-Modified-Since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Correctly invalidate default issuer changes When the default issuer changes, we'll have to mark the invalidation on PR secondary clusters, so they know to update their CRL mapping as well. The swapped issuers will have an updated modification time (which will eventually replicate down and thus be correct), but the CRL modification time is cluster-local information and thus won't be replicated. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * make fmt Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Refactor sendNotModifiedResponseIfNecessary Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation on if-modified-since Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Co-authored-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-08-29 19:28:47 +00:00
}
func TestBackend_InitializeCertificateCounts(t *testing.T) {
t.Parallel()
b, s := CreateBackendWithStorage(t)
ctx := context.Background()
// Set up an Issuer and Role
// We need a root certificate to write/revoke certificates with
resp, err := CBWrite(b, s, "root/generate/internal", map[string]interface{}{
"common_name": "myvault.com",
})
if err != nil {
t.Fatal(err)
}
if resp == nil {
t.Fatal("expected ca info")
}
// Create a role
_, err = CBWrite(b, s, "roles/example", map[string]interface{}{
"allowed_domains": "myvault.com",
"allow_bare_domains": true,
"allow_subdomains": true,
"max_ttl": "2h",
})
if err != nil {
t.Fatal(err)
}
// Put certificates A, B, C, D, E in backend
var certificates []string = []string{"a", "b", "c", "d", "e"}
serials := make([]string, 5)
for i, cn := range certificates {
resp, err = CBWrite(b, s, "issue/example", map[string]interface{}{
"common_name": cn + ".myvault.com",
})
if err != nil {
t.Fatal(err)
}
serials[i] = resp.Data["serial_number"].(string)
}
// Revoke certificates A + B
revocations := serials[0:2]
for _, key := range revocations {
resp, err = CBWrite(b, s, "revoke", map[string]interface{}{
"serial_number": key,
})
if err != nil {
t.Fatal(err)
}
}
// Assert initialize from clean is correct:
b.initializeStoredCertificateCounts(ctx)
if atomic.LoadUint32(b.certCount) != 6 {
t.Fatalf("Failed to count six certificates root,A,B,C,D,E, instead counted %d certs", atomic.LoadUint32(b.certCount))
}
if atomic.LoadUint32(b.revokedCertCount) != 2 {
t.Fatalf("Failed to count two revoked certificates A+B, instead counted %d certs", atomic.LoadUint32(b.revokedCertCount))
}
// Simulates listing while initialize in progress, by "restarting it"
atomic.StoreUint32(b.certCount, 0)
atomic.StoreUint32(b.revokedCertCount, 0)
b.certsCounted.Store(false)
// Revoke certificates C, D
dirtyRevocations := serials[2:4]
for _, key := range dirtyRevocations {
resp, err = CBWrite(b, s, "revoke", map[string]interface{}{
"serial_number": key,
})
if err != nil {
t.Fatal(err)
}
}
// Put certificates F, G in the backend
dirtyCertificates := []string{"f", "g"}
for _, cn := range dirtyCertificates {
resp, err = CBWrite(b, s, "issue/example", map[string]interface{}{
"common_name": cn + ".myvault.com",
})
if err != nil {
t.Fatal(err)
}
}
// Run initialize
b.initializeStoredCertificateCounts(ctx)
// Test certificate count
if *(b.certCount) != 8 {
t.Fatalf("Failed to initialize count of certificates root, A,B,C,D,E,F,G counted %d certs", *(b.certCount))
}
if *(b.revokedCertCount) != 4 {
t.Fatalf("Failed to count revoked certificates A,B,C,D counted %d certs", *(b.revokedCertCount))
}
return
}
// Verify that our default values are consistent when creating an issuer and when we do an
// empty POST update to it. This will hopefully identify if we have different default values
// for fields across the two APIs.
func TestBackend_VerifyIssuerUpdateDefaultsMatchCreation(t *testing.T) {
t.Parallel()
b, s := CreateBackendWithStorage(t)
resp, err := CBWrite(b, s, "root/generate/internal", map[string]interface{}{
"common_name": "myvault.com",
})
requireSuccessNonNilResponse(t, resp, err, "failed generating root issuer")
resp, err = CBRead(b, s, "issuer/default")
requireSuccessNonNilResponse(t, resp, err, "failed reading default issuer")
preUpdateValues := resp.Data
// This field gets reset during issuer update to the empty string
// (meaning Go will auto-detect the rev-sig-algo).
preUpdateValues["revocation_signature_algorithm"] = ""
resp, err = CBWrite(b, s, "issuer/default", map[string]interface{}{})
requireSuccessNonNilResponse(t, resp, err, "failed updating default issuer with no values")
resp, err = CBRead(b, s, "issuer/default")
requireSuccessNonNilResponse(t, resp, err, "failed reading default issuer")
postUpdateValues := resp.Data
require.Equal(t, preUpdateValues, postUpdateValues,
"A value was updated based on the empty update of an issuer, "+
"most likely we have a different set of field parameters across create and update of issuers.")
}
func TestBackend_VerifyPSSKeysIssuersFailImport(t *testing.T) {
t.Parallel()
b, s := CreateBackendWithStorage(t)
// PKCS8 parsing fails on this key due to rsaPSS OID
rsaOIDKey := `
-----BEGIN PRIVATE KEY-----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-----END PRIVATE KEY-----
`
_, err := CBWrite(b, s, "issuers/import/bundle", map[string]interface{}{
"pem_bundle": rsaOIDKey,
})
require.Error(t, err, "expected error importing PKCS8 rsaPSS OID key")
_, err = CBWrite(b, s, "keys/import", map[string]interface{}{
"key": rsaOIDKey,
})
require.Error(t, err, "expected error importing PKCS8 rsaPSS OID key")
// Importing a cert with rsaPSS OID should also fail
rsaOIDCert := `
-----BEGIN CERTIFICATE-----
MIIDfjCCAjGgAwIBAgIBATBCBgkqhkiG9w0BAQowNaAPMA0GCWCGSAFlAwQCAQUA
oRwwGgYJKoZIhvcNAQEIMA0GCWCGSAFlAwQCAQUAogQCAgDeMBMxETAPBgNVBAMM
CHJvb3Qtb2xkMB4XDTIyMDkxNjE0MDEwM1oXDTIzMDkyNjE0MDEwM1owEzERMA8G
A1UEAwwIcm9vdC1vbGQwggEgMAsGCSqGSIb3DQEBCgOCAQ8AMIIBCgKCAQEAtN0/
NPuJHLuyEdBrtUikXoXOV741XZcNvLAIVBIqDA0ege2gXt9A15FGUI4X3u6kT16F
l6MRdtUZ/qNSVs15nK9A1PI/AVekMgTVFTnoCzs550CKN8iRk9Om+lwHimpyXxKk
FW69v8fsXwKEBsz69jjT7HV9VZQ7fQhmE79brAMuwKP1fUQKdHq5OBKtQ7Cl3Gmi
pp0izCsVuQIEkBHvT3UUgyaSp2n+FONpOiyuBoYUH5tVEv9sZzBqSsrYBJYF+Gvf
nFy9AcTdqRe2VX2SjjWjDF84T30OBA798gIFIPwu9R4OjWOlPeh2bo2kGeo3AITj
wFZ28m7kS7kcOtvHpwIDAQABo3UwczAdBgNVHQ4EFgQUVGkTAUJ8inxIVGBlfxf4
cDhRSnowHwYDVR0jBBgwFoAUVGkTAUJ8inxIVGBlfxf4cDhRSnowDAYDVR0TBAUw
AwEB/zAOBgNVHQ8BAf8EBAMCAYYwEwYDVR0lBAwwCgYIKwYBBQUHAwEwQgYJKoZI
hvcNAQEKMDWgDzANBglghkgBZQMEAgEFAKEcMBoGCSqGSIb3DQEBCDANBglghkgB
ZQMEAgEFAKIEAgIA3gOCAQEAQZ3iQ3NjvS4FYJ5WG41huZI0dkvNFNan+ZYWlYHJ
MIQhbFogb/UQB0rlsuldG0+HF1RDXoYNuThfzt5hiBWYEtMBNurezvnOn4DF0hrl
Uk3sBVnvTalVXg+UVjqh9hBGB75JYJl6a5Oa2Zrq++4qGNwjd0FqgnoXzqS5UGuB
TJL8nlnXPuOIK3VHoXEy7l9GtvEzKcys0xa7g1PYpaJ5D2kpbBJmuQGmU6CDcbP+
m0hI4QDfVfHtnBp2VMCvhj0yzowtwF4BFIhv4EXZBU10mzxVj0zyKKft9++X8auH
nebuK22ZwzbPe4NhOvAdfNDElkrrtGvTnzkDB7ezPYjelA==
-----END CERTIFICATE-----
`
_, err = CBWrite(b, s, "issuers/import/bundle", map[string]interface{}{
"pem_bundle": rsaOIDCert,
})
require.Error(t, err, "expected error importing PKCS8 rsaPSS OID cert")
_, err = CBWrite(b, s, "issuers/import/bundle", map[string]interface{}{
"pem_bundle": rsaOIDKey + "\n" + rsaOIDCert,
})
require.Error(t, err, "expected error importing PKCS8 rsaPSS OID key+cert")
_, err = CBWrite(b, s, "issuers/import/bundle", map[string]interface{}{
"pem_bundle": rsaOIDCert + "\n" + rsaOIDKey,
})
require.Error(t, err, "expected error importing PKCS8 rsaPSS OID cert+key")
// After all these errors, we should have zero issuers and keys.
resp, err := CBList(b, s, "issuers")
require.NoError(t, err)
require.Equal(t, nil, resp.Data["keys"])
resp, err = CBList(b, s, "keys")
require.NoError(t, err)
require.Equal(t, nil, resp.Data["keys"])
// If we create a new PSS root, we should be able to issue an intermediate
// under it.
resp, err = CBWrite(b, s, "root/generate/exported", map[string]interface{}{
"use_pss": "true",
"common_name": "root x1 - pss",
"key_type": "ec",
})
require.NoError(t, err)
require.NotNil(t, resp)
require.NotNil(t, resp.Data)
require.NotEmpty(t, resp.Data["certificate"])
require.NotEmpty(t, resp.Data["private_key"])
resp, err = CBWrite(b, s, "intermediate/generate/exported", map[string]interface{}{
"use_pss": "true",
"common_name": "int x1 - pss",
"key_type": "ec",
})
require.NoError(t, err)
require.NotNil(t, resp)
require.NotNil(t, resp.Data)
require.NotEmpty(t, resp.Data["csr"])
require.NotEmpty(t, resp.Data["private_key"])
resp, err = CBWrite(b, s, "issuer/default/sign-intermediate", map[string]interface{}{
"use_pss": "true",
"common_name": "int x1 - pss",
"csr": resp.Data["csr"].(string),
})
require.NoError(t, err)
require.NotNil(t, resp)
require.NotNil(t, resp.Data)
require.NotEmpty(t, resp.Data["certificate"])
resp, err = CBWrite(b, s, "issuers/import/bundle", map[string]interface{}{
"pem_bundle": resp.Data["certificate"].(string),
})
require.NoError(t, err)
// Finally, if we were to take an rsaPSS OID'd CSR and use it against this
// mount, it will fail.
_, err = CBWrite(b, s, "roles/testing", map[string]interface{}{
"allow_any_name": true,
"ttl": "85s",
"key_type": "any",
})
require.NoError(t, err)
// Issuing a leaf from a CSR with rsaPSS OID should fail...
rsaOIDCSR := `-----BEGIN CERTIFICATE REQUEST-----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-----END CERTIFICATE REQUEST-----`
_, err = CBWrite(b, s, "issuer/default/sign/testing", map[string]interface{}{
"common_name": "example.com",
"csr": rsaOIDCSR,
})
require.Error(t, err)
_, err = CBWrite(b, s, "issuer/default/sign-verbatim", map[string]interface{}{
"common_name": "example.com",
"use_pss": true,
"csr": rsaOIDCSR,
})
require.Error(t, err)
_, err = CBWrite(b, s, "issuer/default/sign-intermediate", map[string]interface{}{
"common_name": "faulty x1 - pss",
"use_pss": true,
"csr": rsaOIDCSR,
})
require.Error(t, err)
// Vault has a weird API for signing self-signed certificates. Ensure
// that doesn't accept rsaPSS OID'd certificates either.
_, err = CBWrite(b, s, "issuer/default/sign-self-issued", map[string]interface{}{
"use_pss": true,
"certificate": rsaOIDCert,
})
require.Error(t, err)
// Issuing a regular leaf should succeed.
_, err = CBWrite(b, s, "roles/testing", map[string]interface{}{
"allow_any_name": true,
"ttl": "85s",
"key_type": "rsa",
"use_pss": "true",
})
require.NoError(t, err)
resp, err = CBWrite(b, s, "issuer/default/issue/testing", map[string]interface{}{
"common_name": "example.com",
"use_pss": "true",
})
requireSuccessNonNilResponse(t, resp, err, "failed to issue PSS leaf")
}
func TestPKI_EmptyCRLConfigUpgraded(t *testing.T) {
t.Parallel()
b, s := CreateBackendWithStorage(t)
// Write an empty CRLConfig into storage.
crlConfigEntry, err := logical.StorageEntryJSON("config/crl", &crlConfig{})
require.NoError(t, err)
err = s.Put(ctx, crlConfigEntry)
require.NoError(t, err)
resp, err := CBRead(b, s, "config/crl")
require.NoError(t, err)
require.NotNil(t, resp)
require.NotNil(t, resp.Data)
require.Equal(t, resp.Data["expiry"], defaultCrlConfig.Expiry)
require.Equal(t, resp.Data["disable"], defaultCrlConfig.Disable)
require.Equal(t, resp.Data["ocsp_disable"], defaultCrlConfig.OcspDisable)
require.Equal(t, resp.Data["auto_rebuild"], defaultCrlConfig.AutoRebuild)
require.Equal(t, resp.Data["auto_rebuild_grace_period"], defaultCrlConfig.AutoRebuildGracePeriod)
require.Equal(t, resp.Data["enable_delta"], defaultCrlConfig.EnableDelta)
require.Equal(t, resp.Data["delta_rebuild_interval"], defaultCrlConfig.DeltaRebuildInterval)
}
func TestPKI_ListRevokedCerts(t *testing.T) {
t.Parallel()
b, s := CreateBackendWithStorage(t)
// Test empty cluster
resp, err := CBList(b, s, "certs/revoked")
requireSuccessNonNilResponse(t, resp, err, "failed listing empty cluster")
require.Empty(t, resp.Data, "response map contained data that we did not expect")
// Set up a mount that we can revoke under (We will create 3 leaf certs, 2 of which will be revoked)
resp, err = CBWrite(b, s, "root/generate/internal", map[string]interface{}{
"common_name": "test.com",
"key_type": "ec",
})
requireSuccessNonNilResponse(t, resp, err, "error generating root CA")
requireFieldsSetInResp(t, resp, "serial_number")
issuerSerial := resp.Data["serial_number"]
resp, err = CBWrite(b, s, "roles/test", map[string]interface{}{
"allowed_domains": "test.com",
"allow_subdomains": "true",
"max_ttl": "1h",
})
requireSuccessNonNilResponse(t, resp, err, "error setting up pki role")
resp, err = CBWrite(b, s, "issue/test", map[string]interface{}{
"common_name": "test1.test.com",
})
requireSuccessNonNilResponse(t, resp, err, "error issuing cert 1")
requireFieldsSetInResp(t, resp, "serial_number")
serial1 := resp.Data["serial_number"]
resp, err = CBWrite(b, s, "issue/test", map[string]interface{}{
"common_name": "test2.test.com",
})
requireSuccessNonNilResponse(t, resp, err, "error issuing cert 2")
requireFieldsSetInResp(t, resp, "serial_number")
serial2 := resp.Data["serial_number"]
resp, err = CBWrite(b, s, "issue/test", map[string]interface{}{
"common_name": "test3.test.com",
})
requireSuccessNonNilResponse(t, resp, err, "error issuing cert 2")
requireFieldsSetInResp(t, resp, "serial_number")
serial3 := resp.Data["serial_number"]
resp, err = CBWrite(b, s, "revoke", map[string]interface{}{"serial_number": serial1})
requireSuccessNonNilResponse(t, resp, err, "error revoking cert 1")
resp, err = CBWrite(b, s, "revoke", map[string]interface{}{"serial_number": serial2})
requireSuccessNonNilResponse(t, resp, err, "error revoking cert 2")
// Test that we get back the expected revoked serial numbers.
resp, err = CBList(b, s, "certs/revoked")
requireSuccessNonNilResponse(t, resp, err, "failed listing revoked certs")
requireFieldsSetInResp(t, resp, "keys")
revokedKeys := resp.Data["keys"].([]string)
require.Contains(t, revokedKeys, serial1)
require.Contains(t, revokedKeys, serial2)
require.Equal(t, 2, len(revokedKeys), "Expected 2 revoked entries got %d: %v", len(revokedKeys), revokedKeys)
// Test that listing our certs returns a different response
resp, err = CBList(b, s, "certs")
requireSuccessNonNilResponse(t, resp, err, "failed listing written certs")
requireFieldsSetInResp(t, resp, "keys")
certKeys := resp.Data["keys"].([]string)
require.Contains(t, certKeys, serial1)
require.Contains(t, certKeys, serial2)
require.Contains(t, certKeys, serial3)
require.Contains(t, certKeys, issuerSerial)
require.Equal(t, 4, len(certKeys), "Expected 4 cert entries got %d: %v", len(certKeys), certKeys)
}
Allow templating cluster-local AIA URIs (#18199) * Allow templating of cluster-local AIA URIs This adds a new configuration path, /config/cluster, which retains cluster-local configuration. By extending /config/urls and its issuer counterpart to include an enable_templating parameter, we can allow operators to correctly identify the particular cluster a cert was issued on, and tie its AIA information to this (cluster, issuer) pair dynamically. Notably, this does not solve all usage issues around AIA URIs: the CRL and OCSP responder remain local, meaning that some merge capability is required prior to passing it to other systems if they use CRL files and must validate requests with certs from any arbitrary PR cluster. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation about templated AIAs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * AIA URIs -> AIA URLs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * issuer.AIAURIs might be nil Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow non-nil response to config/urls Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Always validate URLs on config update Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Ensure URLs lack templating parameters Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Review feedback Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-12-05 15:38:26 +00:00
func TestPKI_TemplatedAIAs(t *testing.T) {
t.Parallel()
b, s := CreateBackendWithStorage(t)
// Setting templated AIAs should succeed.
_, err := CBWrite(b, s, "config/cluster", map[string]interface{}{
"path": "http://localhost:8200/v1/pki",
"aia_path": "http://localhost:8200/cdn/pki",
Allow templating cluster-local AIA URIs (#18199) * Allow templating of cluster-local AIA URIs This adds a new configuration path, /config/cluster, which retains cluster-local configuration. By extending /config/urls and its issuer counterpart to include an enable_templating parameter, we can allow operators to correctly identify the particular cluster a cert was issued on, and tie its AIA information to this (cluster, issuer) pair dynamically. Notably, this does not solve all usage issues around AIA URIs: the CRL and OCSP responder remain local, meaning that some merge capability is required prior to passing it to other systems if they use CRL files and must validate requests with certs from any arbitrary PR cluster. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation about templated AIAs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * AIA URIs -> AIA URLs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * issuer.AIAURIs might be nil Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow non-nil response to config/urls Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Always validate URLs on config update Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Ensure URLs lack templating parameters Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Review feedback Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-12-05 15:38:26 +00:00
})
require.NoError(t, err)
aiaData := map[string]interface{}{
"crl_distribution_points": "{{cluster_path}}/issuer/{{issuer_id}}/crl/der",
"issuing_certificates": "{{cluster_aia_path}}/issuer/{{issuer_id}}/der",
Allow templating cluster-local AIA URIs (#18199) * Allow templating of cluster-local AIA URIs This adds a new configuration path, /config/cluster, which retains cluster-local configuration. By extending /config/urls and its issuer counterpart to include an enable_templating parameter, we can allow operators to correctly identify the particular cluster a cert was issued on, and tie its AIA information to this (cluster, issuer) pair dynamically. Notably, this does not solve all usage issues around AIA URIs: the CRL and OCSP responder remain local, meaning that some merge capability is required prior to passing it to other systems if they use CRL files and must validate requests with certs from any arbitrary PR cluster. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation about templated AIAs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * AIA URIs -> AIA URLs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * issuer.AIAURIs might be nil Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow non-nil response to config/urls Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Always validate URLs on config update Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Ensure URLs lack templating parameters Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Review feedback Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-12-05 15:38:26 +00:00
"ocsp_servers": "{{cluster_path}}/ocsp",
"enable_templating": true,
}
_, err = CBWrite(b, s, "config/urls", aiaData)
require.NoError(t, err)
// But root generation will fail.
rootData := map[string]interface{}{
"common_name": "Long-Lived Root X1",
"issuer_name": "long-root-x1",
"key_type": "ec",
}
_, err = CBWrite(b, s, "root/generate/internal", rootData)
require.Error(t, err)
require.Contains(t, err.Error(), "unable to parse AIA URL")
// Clearing the config and regenerating the root should succeed.
_, err = CBWrite(b, s, "config/urls", map[string]interface{}{
"crl_distribution_points": "",
"issuing_certificates": "",
"ocsp_servers": "",
"enable_templating": false,
})
require.NoError(t, err)
resp, err := CBWrite(b, s, "root/generate/internal", rootData)
requireSuccessNonNilResponse(t, resp, err)
issuerId := string(resp.Data["issuer_id"].(issuerID))
// Now write the original AIA config and sign a leaf.
_, err = CBWrite(b, s, "config/urls", aiaData)
require.NoError(t, err)
_, err = CBWrite(b, s, "roles/testing", map[string]interface{}{
"allow_any_name": "true",
"key_type": "ec",
"ttl": "50m",
})
require.NoError(t, err)
resp, err = CBWrite(b, s, "issue/testing", map[string]interface{}{
"common_name": "example.com",
})
requireSuccessNonNilResponse(t, resp, err)
// Validate the AIA info is correctly templated.
cert := parseCert(t, resp.Data["certificate"].(string))
require.Equal(t, cert.OCSPServer, []string{"http://localhost:8200/v1/pki/ocsp"})
require.Equal(t, cert.IssuingCertificateURL, []string{"http://localhost:8200/cdn/pki/issuer/" + issuerId + "/der"})
Allow templating cluster-local AIA URIs (#18199) * Allow templating of cluster-local AIA URIs This adds a new configuration path, /config/cluster, which retains cluster-local configuration. By extending /config/urls and its issuer counterpart to include an enable_templating parameter, we can allow operators to correctly identify the particular cluster a cert was issued on, and tie its AIA information to this (cluster, issuer) pair dynamically. Notably, this does not solve all usage issues around AIA URIs: the CRL and OCSP responder remain local, meaning that some merge capability is required prior to passing it to other systems if they use CRL files and must validate requests with certs from any arbitrary PR cluster. Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add documentation about templated AIAs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add changelog entry Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Add tests Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * AIA URIs -> AIA URLs Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * issuer.AIAURIs might be nil Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Allow non-nil response to config/urls Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Always validate URLs on config update Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Ensure URLs lack templating parameters Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> * Review feedback Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com> Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
2022-12-05 15:38:26 +00:00
require.Equal(t, cert.CRLDistributionPoints, []string{"http://localhost:8200/v1/pki/issuer/" + issuerId + "/crl/der"})
// Modify our issuer to set custom AIAs: these URLs are bad.
_, err = CBPatch(b, s, "issuer/default", map[string]interface{}{
"enable_aia_url_templating": "false",
"crl_distribution_points": "a",
"issuing_certificates": "b",
"ocsp_servers": "c",
})
require.Error(t, err)
// These URLs are good.
_, err = CBPatch(b, s, "issuer/default", map[string]interface{}{
"enable_aia_url_templating": "false",
"crl_distribution_points": "http://localhost/a",
"issuing_certificates": "http://localhost/b",
"ocsp_servers": "http://localhost/c",
})
resp, err = CBWrite(b, s, "issue/testing", map[string]interface{}{
"common_name": "example.com",
})
requireSuccessNonNilResponse(t, resp, err)
// Validate the AIA info is correctly templated.
cert = parseCert(t, resp.Data["certificate"].(string))
require.Equal(t, cert.OCSPServer, []string{"http://localhost/c"})
require.Equal(t, cert.IssuingCertificateURL, []string{"http://localhost/b"})
require.Equal(t, cert.CRLDistributionPoints, []string{"http://localhost/a"})
// These URLs are bad, but will fail at issuance time due to AIA templating.
resp, err = CBPatch(b, s, "issuer/default", map[string]interface{}{
"enable_aia_url_templating": "true",
"crl_distribution_points": "a",
"issuing_certificates": "b",
"ocsp_servers": "c",
})
requireSuccessNonNilResponse(t, resp, err)
require.NotEmpty(t, resp.Warnings)
_, err = CBWrite(b, s, "issue/testing", map[string]interface{}{
"common_name": "example.com",
})
require.Error(t, err)
}
2018-02-20 05:03:45 +00:00
var (
initTest sync.Once
rsaCAKey string
rsaCACert string
ecCAKey string
ecCACert string
edCAKey string
edCACert string
)