2014-08-06 22:08:17 +00:00
|
|
|
package acl
|
|
|
|
|
|
|
|
import (
|
|
|
|
"fmt"
|
2014-12-01 03:18:16 +00:00
|
|
|
|
2017-09-14 19:31:01 +00:00
|
|
|
"github.com/hashicorp/consul/sentinel"
|
2014-08-06 22:08:17 +00:00
|
|
|
"github.com/hashicorp/hcl"
|
|
|
|
)
|
|
|
|
|
|
|
|
const (
|
Creates new "prepared-query" ACL type and new token capture behavior.
Prior to this change, prepared queries had the following behavior for
ACLs, which will need to change to support templates:
1. A management token, or a token with read access to the service being
queried needed to be provided in order to create a prepared query.
2. The token used to create the prepared query was stored with the query
in the state store and used to execute the query.
3. A management token, or the token used to create the query needed to be
supplied to perform and CRUD operations on an existing prepared query.
This was pretty subtle and complicated behavior, and won't work for
templates since the service name is computed at execution time. To solve
this, we introduce a new "prepared-query" ACL type, where the prefix
applies to the query name for static prepared query types and to the
prefix for template prepared query types.
With this change, the new behavior is:
1. A management token, or a token with "prepared-query" write access to
the query name or (soon) the given template prefix is required to do
any CRUD operations on a prepared query, or to list prepared queries
(the list is filtered by this ACL).
2. You will no longer need a management token to list prepared queries,
but you will only be able to see prepared queries that you have access
to (you get an empty list instead of permission denied).
3. When listing or getting a query, because it was easy to capture
management tokens given the past behavior, this will always blank out
the "Token" field (replacing the contents as <hidden>) for all tokens
unless a management token is supplied. Going forward, we should
discourage people from binding tokens for execution unless strictly
necessary.
4. No token will be captured by default when a prepared query is created.
If the user wishes to supply an execution token then can pass it in via
the "Token" field in the prepared query definition. Otherwise, this
field will default to empty.
5. At execution time, we will use the captured token if it exists with the
prepared query definition, otherwise we will use the token that's passed
in with the request, just like we do for other RPCs (or you can use the
agent's configured token for DNS).
6. Prepared queries with no name (accessible only by ID) will not require
ACLs to create or modify (execution time will depend on the service ACL
configuration). Our argument here is that these are designed to be
ephemeral and the IDs are as good as an ACL. Management tokens will be
able to list all of these.
These changes enable templates, but also enable delegation of authority to
manage the prepared query namespace.
2016-02-23 08:12:58 +00:00
|
|
|
PolicyDeny = "deny"
|
|
|
|
PolicyRead = "read"
|
|
|
|
PolicyWrite = "write"
|
2017-10-02 22:10:21 +00:00
|
|
|
PolicyList = "list"
|
2014-08-06 22:08:17 +00:00
|
|
|
)
|
|
|
|
|
|
|
|
// Policy is used to represent the policy specified by
|
|
|
|
// an ACL configuration.
|
|
|
|
type Policy struct {
|
Creates new "prepared-query" ACL type and new token capture behavior.
Prior to this change, prepared queries had the following behavior for
ACLs, which will need to change to support templates:
1. A management token, or a token with read access to the service being
queried needed to be provided in order to create a prepared query.
2. The token used to create the prepared query was stored with the query
in the state store and used to execute the query.
3. A management token, or the token used to create the query needed to be
supplied to perform and CRUD operations on an existing prepared query.
This was pretty subtle and complicated behavior, and won't work for
templates since the service name is computed at execution time. To solve
this, we introduce a new "prepared-query" ACL type, where the prefix
applies to the query name for static prepared query types and to the
prefix for template prepared query types.
With this change, the new behavior is:
1. A management token, or a token with "prepared-query" write access to
the query name or (soon) the given template prefix is required to do
any CRUD operations on a prepared query, or to list prepared queries
(the list is filtered by this ACL).
2. You will no longer need a management token to list prepared queries,
but you will only be able to see prepared queries that you have access
to (you get an empty list instead of permission denied).
3. When listing or getting a query, because it was easy to capture
management tokens given the past behavior, this will always blank out
the "Token" field (replacing the contents as <hidden>) for all tokens
unless a management token is supplied. Going forward, we should
discourage people from binding tokens for execution unless strictly
necessary.
4. No token will be captured by default when a prepared query is created.
If the user wishes to supply an execution token then can pass it in via
the "Token" field in the prepared query definition. Otherwise, this
field will default to empty.
5. At execution time, we will use the captured token if it exists with the
prepared query definition, otherwise we will use the token that's passed
in with the request, just like we do for other RPCs (or you can use the
agent's configured token for DNS).
6. Prepared queries with no name (accessible only by ID) will not require
ACLs to create or modify (execution time will depend on the service ACL
configuration). Our argument here is that these are designed to be
ephemeral and the IDs are as good as an ACL. Management tokens will be
able to list all of these.
These changes enable templates, but also enable delegation of authority to
manage the prepared query namespace.
2016-02-23 08:12:58 +00:00
|
|
|
ID string `hcl:"-"`
|
2016-12-13 07:05:11 +00:00
|
|
|
Agents []*AgentPolicy `hcl:"agent,expand"`
|
Creates new "prepared-query" ACL type and new token capture behavior.
Prior to this change, prepared queries had the following behavior for
ACLs, which will need to change to support templates:
1. A management token, or a token with read access to the service being
queried needed to be provided in order to create a prepared query.
2. The token used to create the prepared query was stored with the query
in the state store and used to execute the query.
3. A management token, or the token used to create the query needed to be
supplied to perform and CRUD operations on an existing prepared query.
This was pretty subtle and complicated behavior, and won't work for
templates since the service name is computed at execution time. To solve
this, we introduce a new "prepared-query" ACL type, where the prefix
applies to the query name for static prepared query types and to the
prefix for template prepared query types.
With this change, the new behavior is:
1. A management token, or a token with "prepared-query" write access to
the query name or (soon) the given template prefix is required to do
any CRUD operations on a prepared query, or to list prepared queries
(the list is filtered by this ACL).
2. You will no longer need a management token to list prepared queries,
but you will only be able to see prepared queries that you have access
to (you get an empty list instead of permission denied).
3. When listing or getting a query, because it was easy to capture
management tokens given the past behavior, this will always blank out
the "Token" field (replacing the contents as <hidden>) for all tokens
unless a management token is supplied. Going forward, we should
discourage people from binding tokens for execution unless strictly
necessary.
4. No token will be captured by default when a prepared query is created.
If the user wishes to supply an execution token then can pass it in via
the "Token" field in the prepared query definition. Otherwise, this
field will default to empty.
5. At execution time, we will use the captured token if it exists with the
prepared query definition, otherwise we will use the token that's passed
in with the request, just like we do for other RPCs (or you can use the
agent's configured token for DNS).
6. Prepared queries with no name (accessible only by ID) will not require
ACLs to create or modify (execution time will depend on the service ACL
configuration). Our argument here is that these are designed to be
ephemeral and the IDs are as good as an ACL. Management tokens will be
able to list all of these.
These changes enable templates, but also enable delegation of authority to
manage the prepared query namespace.
2016-02-23 08:12:58 +00:00
|
|
|
Keys []*KeyPolicy `hcl:"key,expand"`
|
2016-12-07 04:05:15 +00:00
|
|
|
Nodes []*NodePolicy `hcl:"node,expand"`
|
Creates new "prepared-query" ACL type and new token capture behavior.
Prior to this change, prepared queries had the following behavior for
ACLs, which will need to change to support templates:
1. A management token, or a token with read access to the service being
queried needed to be provided in order to create a prepared query.
2. The token used to create the prepared query was stored with the query
in the state store and used to execute the query.
3. A management token, or the token used to create the query needed to be
supplied to perform and CRUD operations on an existing prepared query.
This was pretty subtle and complicated behavior, and won't work for
templates since the service name is computed at execution time. To solve
this, we introduce a new "prepared-query" ACL type, where the prefix
applies to the query name for static prepared query types and to the
prefix for template prepared query types.
With this change, the new behavior is:
1. A management token, or a token with "prepared-query" write access to
the query name or (soon) the given template prefix is required to do
any CRUD operations on a prepared query, or to list prepared queries
(the list is filtered by this ACL).
2. You will no longer need a management token to list prepared queries,
but you will only be able to see prepared queries that you have access
to (you get an empty list instead of permission denied).
3. When listing or getting a query, because it was easy to capture
management tokens given the past behavior, this will always blank out
the "Token" field (replacing the contents as <hidden>) for all tokens
unless a management token is supplied. Going forward, we should
discourage people from binding tokens for execution unless strictly
necessary.
4. No token will be captured by default when a prepared query is created.
If the user wishes to supply an execution token then can pass it in via
the "Token" field in the prepared query definition. Otherwise, this
field will default to empty.
5. At execution time, we will use the captured token if it exists with the
prepared query definition, otherwise we will use the token that's passed
in with the request, just like we do for other RPCs (or you can use the
agent's configured token for DNS).
6. Prepared queries with no name (accessible only by ID) will not require
ACLs to create or modify (execution time will depend on the service ACL
configuration). Our argument here is that these are designed to be
ephemeral and the IDs are as good as an ACL. Management tokens will be
able to list all of these.
These changes enable templates, but also enable delegation of authority to
manage the prepared query namespace.
2016-02-23 08:12:58 +00:00
|
|
|
Services []*ServicePolicy `hcl:"service,expand"`
|
2016-12-13 04:20:28 +00:00
|
|
|
Sessions []*SessionPolicy `hcl:"session,expand"`
|
Creates new "prepared-query" ACL type and new token capture behavior.
Prior to this change, prepared queries had the following behavior for
ACLs, which will need to change to support templates:
1. A management token, or a token with read access to the service being
queried needed to be provided in order to create a prepared query.
2. The token used to create the prepared query was stored with the query
in the state store and used to execute the query.
3. A management token, or the token used to create the query needed to be
supplied to perform and CRUD operations on an existing prepared query.
This was pretty subtle and complicated behavior, and won't work for
templates since the service name is computed at execution time. To solve
this, we introduce a new "prepared-query" ACL type, where the prefix
applies to the query name for static prepared query types and to the
prefix for template prepared query types.
With this change, the new behavior is:
1. A management token, or a token with "prepared-query" write access to
the query name or (soon) the given template prefix is required to do
any CRUD operations on a prepared query, or to list prepared queries
(the list is filtered by this ACL).
2. You will no longer need a management token to list prepared queries,
but you will only be able to see prepared queries that you have access
to (you get an empty list instead of permission denied).
3. When listing or getting a query, because it was easy to capture
management tokens given the past behavior, this will always blank out
the "Token" field (replacing the contents as <hidden>) for all tokens
unless a management token is supplied. Going forward, we should
discourage people from binding tokens for execution unless strictly
necessary.
4. No token will be captured by default when a prepared query is created.
If the user wishes to supply an execution token then can pass it in via
the "Token" field in the prepared query definition. Otherwise, this
field will default to empty.
5. At execution time, we will use the captured token if it exists with the
prepared query definition, otherwise we will use the token that's passed
in with the request, just like we do for other RPCs (or you can use the
agent's configured token for DNS).
6. Prepared queries with no name (accessible only by ID) will not require
ACLs to create or modify (execution time will depend on the service ACL
configuration). Our argument here is that these are designed to be
ephemeral and the IDs are as good as an ACL. Management tokens will be
able to list all of these.
These changes enable templates, but also enable delegation of authority to
manage the prepared query namespace.
2016-02-23 08:12:58 +00:00
|
|
|
Events []*EventPolicy `hcl:"event,expand"`
|
2016-02-25 00:57:55 +00:00
|
|
|
PreparedQueries []*PreparedQueryPolicy `hcl:"query,expand"`
|
Creates new "prepared-query" ACL type and new token capture behavior.
Prior to this change, prepared queries had the following behavior for
ACLs, which will need to change to support templates:
1. A management token, or a token with read access to the service being
queried needed to be provided in order to create a prepared query.
2. The token used to create the prepared query was stored with the query
in the state store and used to execute the query.
3. A management token, or the token used to create the query needed to be
supplied to perform and CRUD operations on an existing prepared query.
This was pretty subtle and complicated behavior, and won't work for
templates since the service name is computed at execution time. To solve
this, we introduce a new "prepared-query" ACL type, where the prefix
applies to the query name for static prepared query types and to the
prefix for template prepared query types.
With this change, the new behavior is:
1. A management token, or a token with "prepared-query" write access to
the query name or (soon) the given template prefix is required to do
any CRUD operations on a prepared query, or to list prepared queries
(the list is filtered by this ACL).
2. You will no longer need a management token to list prepared queries,
but you will only be able to see prepared queries that you have access
to (you get an empty list instead of permission denied).
3. When listing or getting a query, because it was easy to capture
management tokens given the past behavior, this will always blank out
the "Token" field (replacing the contents as <hidden>) for all tokens
unless a management token is supplied. Going forward, we should
discourage people from binding tokens for execution unless strictly
necessary.
4. No token will be captured by default when a prepared query is created.
If the user wishes to supply an execution token then can pass it in via
the "Token" field in the prepared query definition. Otherwise, this
field will default to empty.
5. At execution time, we will use the captured token if it exists with the
prepared query definition, otherwise we will use the token that's passed
in with the request, just like we do for other RPCs (or you can use the
agent's configured token for DNS).
6. Prepared queries with no name (accessible only by ID) will not require
ACLs to create or modify (execution time will depend on the service ACL
configuration). Our argument here is that these are designed to be
ephemeral and the IDs are as good as an ACL. Management tokens will be
able to list all of these.
These changes enable templates, but also enable delegation of authority to
manage the prepared query namespace.
2016-02-23 08:12:58 +00:00
|
|
|
Keyring string `hcl:"keyring"`
|
2016-08-30 02:09:57 +00:00
|
|
|
Operator string `hcl:"operator"`
|
2014-08-06 22:08:17 +00:00
|
|
|
}
|
|
|
|
|
2017-09-14 19:31:01 +00:00
|
|
|
// Sentinel defines a snippet of Sentinel code that can be attached to a policy.
|
|
|
|
type Sentinel struct {
|
|
|
|
Code string
|
|
|
|
EnforcementLevel string
|
|
|
|
}
|
|
|
|
|
2016-12-13 07:05:11 +00:00
|
|
|
// AgentPolicy represents a policy for working with agent endpoints on nodes
|
|
|
|
// with specific name prefixes.
|
|
|
|
type AgentPolicy struct {
|
|
|
|
Node string `hcl:",key"`
|
|
|
|
Policy string
|
|
|
|
}
|
|
|
|
|
|
|
|
func (a *AgentPolicy) GoString() string {
|
|
|
|
return fmt.Sprintf("%#v", *a)
|
|
|
|
}
|
|
|
|
|
2014-08-06 22:08:17 +00:00
|
|
|
// KeyPolicy represents a policy for a key
|
|
|
|
type KeyPolicy struct {
|
2017-09-14 19:31:01 +00:00
|
|
|
Prefix string `hcl:",key"`
|
|
|
|
Policy string
|
|
|
|
Sentinel Sentinel
|
2014-08-08 22:57:28 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
func (k *KeyPolicy) GoString() string {
|
|
|
|
return fmt.Sprintf("%#v", *k)
|
2014-08-06 22:08:17 +00:00
|
|
|
}
|
|
|
|
|
2016-12-07 04:05:15 +00:00
|
|
|
// NodePolicy represents a policy for a node
|
|
|
|
type NodePolicy struct {
|
2017-09-14 19:31:01 +00:00
|
|
|
Name string `hcl:",key"`
|
|
|
|
Policy string
|
|
|
|
Sentinel Sentinel
|
2016-12-07 04:05:15 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
func (n *NodePolicy) GoString() string {
|
|
|
|
return fmt.Sprintf("%#v", *n)
|
|
|
|
}
|
|
|
|
|
2014-12-01 03:18:16 +00:00
|
|
|
// ServicePolicy represents a policy for a service
|
|
|
|
type ServicePolicy struct {
|
2017-09-14 19:31:01 +00:00
|
|
|
Name string `hcl:",key"`
|
|
|
|
Policy string
|
|
|
|
Sentinel Sentinel
|
2014-12-01 03:18:16 +00:00
|
|
|
}
|
|
|
|
|
2016-12-07 04:05:15 +00:00
|
|
|
func (s *ServicePolicy) GoString() string {
|
|
|
|
return fmt.Sprintf("%#v", *s)
|
2014-12-01 03:18:16 +00:00
|
|
|
}
|
|
|
|
|
2016-12-13 04:20:28 +00:00
|
|
|
// SessionPolicy represents a policy for making sessions tied to specific node
|
|
|
|
// name prefixes.
|
|
|
|
type SessionPolicy struct {
|
|
|
|
Node string `hcl:",key"`
|
|
|
|
Policy string
|
|
|
|
}
|
|
|
|
|
|
|
|
func (s *SessionPolicy) GoString() string {
|
|
|
|
return fmt.Sprintf("%#v", *s)
|
|
|
|
}
|
|
|
|
|
2015-06-18 01:56:29 +00:00
|
|
|
// EventPolicy represents a user event policy.
|
|
|
|
type EventPolicy struct {
|
|
|
|
Event string `hcl:",key"`
|
|
|
|
Policy string
|
|
|
|
}
|
|
|
|
|
|
|
|
func (e *EventPolicy) GoString() string {
|
|
|
|
return fmt.Sprintf("%#v", *e)
|
|
|
|
}
|
|
|
|
|
Creates new "prepared-query" ACL type and new token capture behavior.
Prior to this change, prepared queries had the following behavior for
ACLs, which will need to change to support templates:
1. A management token, or a token with read access to the service being
queried needed to be provided in order to create a prepared query.
2. The token used to create the prepared query was stored with the query
in the state store and used to execute the query.
3. A management token, or the token used to create the query needed to be
supplied to perform and CRUD operations on an existing prepared query.
This was pretty subtle and complicated behavior, and won't work for
templates since the service name is computed at execution time. To solve
this, we introduce a new "prepared-query" ACL type, where the prefix
applies to the query name for static prepared query types and to the
prefix for template prepared query types.
With this change, the new behavior is:
1. A management token, or a token with "prepared-query" write access to
the query name or (soon) the given template prefix is required to do
any CRUD operations on a prepared query, or to list prepared queries
(the list is filtered by this ACL).
2. You will no longer need a management token to list prepared queries,
but you will only be able to see prepared queries that you have access
to (you get an empty list instead of permission denied).
3. When listing or getting a query, because it was easy to capture
management tokens given the past behavior, this will always blank out
the "Token" field (replacing the contents as <hidden>) for all tokens
unless a management token is supplied. Going forward, we should
discourage people from binding tokens for execution unless strictly
necessary.
4. No token will be captured by default when a prepared query is created.
If the user wishes to supply an execution token then can pass it in via
the "Token" field in the prepared query definition. Otherwise, this
field will default to empty.
5. At execution time, we will use the captured token if it exists with the
prepared query definition, otherwise we will use the token that's passed
in with the request, just like we do for other RPCs (or you can use the
agent's configured token for DNS).
6. Prepared queries with no name (accessible only by ID) will not require
ACLs to create or modify (execution time will depend on the service ACL
configuration). Our argument here is that these are designed to be
ephemeral and the IDs are as good as an ACL. Management tokens will be
able to list all of these.
These changes enable templates, but also enable delegation of authority to
manage the prepared query namespace.
2016-02-23 08:12:58 +00:00
|
|
|
// PreparedQueryPolicy represents a prepared query policy.
|
|
|
|
type PreparedQueryPolicy struct {
|
|
|
|
Prefix string `hcl:",key"`
|
|
|
|
Policy string
|
|
|
|
}
|
|
|
|
|
2016-12-07 04:05:15 +00:00
|
|
|
func (p *PreparedQueryPolicy) GoString() string {
|
|
|
|
return fmt.Sprintf("%#v", *p)
|
Creates new "prepared-query" ACL type and new token capture behavior.
Prior to this change, prepared queries had the following behavior for
ACLs, which will need to change to support templates:
1. A management token, or a token with read access to the service being
queried needed to be provided in order to create a prepared query.
2. The token used to create the prepared query was stored with the query
in the state store and used to execute the query.
3. A management token, or the token used to create the query needed to be
supplied to perform and CRUD operations on an existing prepared query.
This was pretty subtle and complicated behavior, and won't work for
templates since the service name is computed at execution time. To solve
this, we introduce a new "prepared-query" ACL type, where the prefix
applies to the query name for static prepared query types and to the
prefix for template prepared query types.
With this change, the new behavior is:
1. A management token, or a token with "prepared-query" write access to
the query name or (soon) the given template prefix is required to do
any CRUD operations on a prepared query, or to list prepared queries
(the list is filtered by this ACL).
2. You will no longer need a management token to list prepared queries,
but you will only be able to see prepared queries that you have access
to (you get an empty list instead of permission denied).
3. When listing or getting a query, because it was easy to capture
management tokens given the past behavior, this will always blank out
the "Token" field (replacing the contents as <hidden>) for all tokens
unless a management token is supplied. Going forward, we should
discourage people from binding tokens for execution unless strictly
necessary.
4. No token will be captured by default when a prepared query is created.
If the user wishes to supply an execution token then can pass it in via
the "Token" field in the prepared query definition. Otherwise, this
field will default to empty.
5. At execution time, we will use the captured token if it exists with the
prepared query definition, otherwise we will use the token that's passed
in with the request, just like we do for other RPCs (or you can use the
agent's configured token for DNS).
6. Prepared queries with no name (accessible only by ID) will not require
ACLs to create or modify (execution time will depend on the service ACL
configuration). Our argument here is that these are designed to be
ephemeral and the IDs are as good as an ACL. Management tokens will be
able to list all of these.
These changes enable templates, but also enable delegation of authority to
manage the prepared query namespace.
2016-02-23 08:12:58 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
// isPolicyValid makes sure the given string matches one of the valid policies.
|
|
|
|
func isPolicyValid(policy string) bool {
|
|
|
|
switch policy {
|
|
|
|
case PolicyDeny:
|
|
|
|
return true
|
|
|
|
case PolicyRead:
|
|
|
|
return true
|
|
|
|
case PolicyWrite:
|
|
|
|
return true
|
2017-10-02 22:10:21 +00:00
|
|
|
case PolicyList:
|
|
|
|
return true
|
Creates new "prepared-query" ACL type and new token capture behavior.
Prior to this change, prepared queries had the following behavior for
ACLs, which will need to change to support templates:
1. A management token, or a token with read access to the service being
queried needed to be provided in order to create a prepared query.
2. The token used to create the prepared query was stored with the query
in the state store and used to execute the query.
3. A management token, or the token used to create the query needed to be
supplied to perform and CRUD operations on an existing prepared query.
This was pretty subtle and complicated behavior, and won't work for
templates since the service name is computed at execution time. To solve
this, we introduce a new "prepared-query" ACL type, where the prefix
applies to the query name for static prepared query types and to the
prefix for template prepared query types.
With this change, the new behavior is:
1. A management token, or a token with "prepared-query" write access to
the query name or (soon) the given template prefix is required to do
any CRUD operations on a prepared query, or to list prepared queries
(the list is filtered by this ACL).
2. You will no longer need a management token to list prepared queries,
but you will only be able to see prepared queries that you have access
to (you get an empty list instead of permission denied).
3. When listing or getting a query, because it was easy to capture
management tokens given the past behavior, this will always blank out
the "Token" field (replacing the contents as <hidden>) for all tokens
unless a management token is supplied. Going forward, we should
discourage people from binding tokens for execution unless strictly
necessary.
4. No token will be captured by default when a prepared query is created.
If the user wishes to supply an execution token then can pass it in via
the "Token" field in the prepared query definition. Otherwise, this
field will default to empty.
5. At execution time, we will use the captured token if it exists with the
prepared query definition, otherwise we will use the token that's passed
in with the request, just like we do for other RPCs (or you can use the
agent's configured token for DNS).
6. Prepared queries with no name (accessible only by ID) will not require
ACLs to create or modify (execution time will depend on the service ACL
configuration). Our argument here is that these are designed to be
ephemeral and the IDs are as good as an ACL. Management tokens will be
able to list all of these.
These changes enable templates, but also enable delegation of authority to
manage the prepared query namespace.
2016-02-23 08:12:58 +00:00
|
|
|
default:
|
|
|
|
return false
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2017-09-14 19:31:01 +00:00
|
|
|
// isSentinelValid makes sure the given sentinel block is valid, and will skip
|
|
|
|
// out if the evaluator is nil.
|
|
|
|
func isSentinelValid(sentinel sentinel.Evaluator, basicPolicy string, sp Sentinel) error {
|
|
|
|
// TODO (slackpad) - Need a unit test for this.
|
|
|
|
|
|
|
|
// Sentinel not enabled at all, or for this policy.
|
|
|
|
if sentinel == nil {
|
|
|
|
return nil
|
|
|
|
}
|
|
|
|
if sp.Code == "" {
|
|
|
|
return nil
|
|
|
|
}
|
|
|
|
|
|
|
|
// We only allow sentinel code on write policies at this time.
|
|
|
|
if basicPolicy != PolicyWrite {
|
|
|
|
return fmt.Errorf("code is only allowed for write policies")
|
|
|
|
}
|
|
|
|
|
|
|
|
// Validate the sentinel parts.
|
|
|
|
switch sp.EnforcementLevel {
|
|
|
|
case "", "soft-mandatory", "hard-mandatory":
|
|
|
|
// OK
|
|
|
|
default:
|
|
|
|
return fmt.Errorf("unsupported enforcement level %q", sp.EnforcementLevel)
|
|
|
|
}
|
|
|
|
return sentinel.Compile(sp.Code)
|
|
|
|
}
|
|
|
|
|
2014-08-06 22:08:17 +00:00
|
|
|
// Parse is used to parse the specified ACL rules into an
|
|
|
|
// intermediary set of policies, before being compiled into
|
|
|
|
// the ACL
|
2017-09-14 19:31:01 +00:00
|
|
|
func Parse(rules string, sentinel sentinel.Evaluator) (*Policy, error) {
|
2014-08-06 22:08:17 +00:00
|
|
|
// Decode the rules
|
|
|
|
p := &Policy{}
|
2014-08-08 21:36:09 +00:00
|
|
|
if rules == "" {
|
|
|
|
// Hot path for empty rules
|
|
|
|
return p, nil
|
|
|
|
}
|
|
|
|
|
2014-08-06 22:08:17 +00:00
|
|
|
if err := hcl.Decode(p, rules); err != nil {
|
|
|
|
return nil, fmt.Errorf("Failed to parse ACL rules: %v", err)
|
|
|
|
}
|
|
|
|
|
2016-12-13 07:05:11 +00:00
|
|
|
// Validate the agent policy
|
|
|
|
for _, ap := range p.Agents {
|
|
|
|
if !isPolicyValid(ap.Policy) {
|
|
|
|
return nil, fmt.Errorf("Invalid agent policy: %#v", ap)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-08-06 22:08:17 +00:00
|
|
|
// Validate the key policy
|
|
|
|
for _, kp := range p.Keys {
|
Creates new "prepared-query" ACL type and new token capture behavior.
Prior to this change, prepared queries had the following behavior for
ACLs, which will need to change to support templates:
1. A management token, or a token with read access to the service being
queried needed to be provided in order to create a prepared query.
2. The token used to create the prepared query was stored with the query
in the state store and used to execute the query.
3. A management token, or the token used to create the query needed to be
supplied to perform and CRUD operations on an existing prepared query.
This was pretty subtle and complicated behavior, and won't work for
templates since the service name is computed at execution time. To solve
this, we introduce a new "prepared-query" ACL type, where the prefix
applies to the query name for static prepared query types and to the
prefix for template prepared query types.
With this change, the new behavior is:
1. A management token, or a token with "prepared-query" write access to
the query name or (soon) the given template prefix is required to do
any CRUD operations on a prepared query, or to list prepared queries
(the list is filtered by this ACL).
2. You will no longer need a management token to list prepared queries,
but you will only be able to see prepared queries that you have access
to (you get an empty list instead of permission denied).
3. When listing or getting a query, because it was easy to capture
management tokens given the past behavior, this will always blank out
the "Token" field (replacing the contents as <hidden>) for all tokens
unless a management token is supplied. Going forward, we should
discourage people from binding tokens for execution unless strictly
necessary.
4. No token will be captured by default when a prepared query is created.
If the user wishes to supply an execution token then can pass it in via
the "Token" field in the prepared query definition. Otherwise, this
field will default to empty.
5. At execution time, we will use the captured token if it exists with the
prepared query definition, otherwise we will use the token that's passed
in with the request, just like we do for other RPCs (or you can use the
agent's configured token for DNS).
6. Prepared queries with no name (accessible only by ID) will not require
ACLs to create or modify (execution time will depend on the service ACL
configuration). Our argument here is that these are designed to be
ephemeral and the IDs are as good as an ACL. Management tokens will be
able to list all of these.
These changes enable templates, but also enable delegation of authority to
manage the prepared query namespace.
2016-02-23 08:12:58 +00:00
|
|
|
if !isPolicyValid(kp.Policy) {
|
2014-08-06 22:08:17 +00:00
|
|
|
return nil, fmt.Errorf("Invalid key policy: %#v", kp)
|
|
|
|
}
|
2017-09-14 19:31:01 +00:00
|
|
|
if err := isSentinelValid(sentinel, kp.Policy, kp.Sentinel); err != nil {
|
|
|
|
return nil, fmt.Errorf("Invalid key Sentinel policy: %#v, got error:%v", kp, err)
|
|
|
|
}
|
2014-08-06 22:08:17 +00:00
|
|
|
}
|
2014-12-01 03:18:16 +00:00
|
|
|
|
2016-12-07 04:05:15 +00:00
|
|
|
// Validate the node policies
|
|
|
|
for _, np := range p.Nodes {
|
|
|
|
if !isPolicyValid(np.Policy) {
|
|
|
|
return nil, fmt.Errorf("Invalid node policy: %#v", np)
|
|
|
|
}
|
2017-09-14 19:31:01 +00:00
|
|
|
if err := isSentinelValid(sentinel, np.Policy, np.Sentinel); err != nil {
|
|
|
|
return nil, fmt.Errorf("Invalid node Sentinel policy: %#v, got error:%v", np, err)
|
|
|
|
}
|
2016-12-07 04:05:15 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
// Validate the service policies
|
2014-12-01 03:18:16 +00:00
|
|
|
for _, sp := range p.Services {
|
Creates new "prepared-query" ACL type and new token capture behavior.
Prior to this change, prepared queries had the following behavior for
ACLs, which will need to change to support templates:
1. A management token, or a token with read access to the service being
queried needed to be provided in order to create a prepared query.
2. The token used to create the prepared query was stored with the query
in the state store and used to execute the query.
3. A management token, or the token used to create the query needed to be
supplied to perform and CRUD operations on an existing prepared query.
This was pretty subtle and complicated behavior, and won't work for
templates since the service name is computed at execution time. To solve
this, we introduce a new "prepared-query" ACL type, where the prefix
applies to the query name for static prepared query types and to the
prefix for template prepared query types.
With this change, the new behavior is:
1. A management token, or a token with "prepared-query" write access to
the query name or (soon) the given template prefix is required to do
any CRUD operations on a prepared query, or to list prepared queries
(the list is filtered by this ACL).
2. You will no longer need a management token to list prepared queries,
but you will only be able to see prepared queries that you have access
to (you get an empty list instead of permission denied).
3. When listing or getting a query, because it was easy to capture
management tokens given the past behavior, this will always blank out
the "Token" field (replacing the contents as <hidden>) for all tokens
unless a management token is supplied. Going forward, we should
discourage people from binding tokens for execution unless strictly
necessary.
4. No token will be captured by default when a prepared query is created.
If the user wishes to supply an execution token then can pass it in via
the "Token" field in the prepared query definition. Otherwise, this
field will default to empty.
5. At execution time, we will use the captured token if it exists with the
prepared query definition, otherwise we will use the token that's passed
in with the request, just like we do for other RPCs (or you can use the
agent's configured token for DNS).
6. Prepared queries with no name (accessible only by ID) will not require
ACLs to create or modify (execution time will depend on the service ACL
configuration). Our argument here is that these are designed to be
ephemeral and the IDs are as good as an ACL. Management tokens will be
able to list all of these.
These changes enable templates, but also enable delegation of authority to
manage the prepared query namespace.
2016-02-23 08:12:58 +00:00
|
|
|
if !isPolicyValid(sp.Policy) {
|
2014-12-01 03:18:16 +00:00
|
|
|
return nil, fmt.Errorf("Invalid service policy: %#v", sp)
|
|
|
|
}
|
2017-09-14 19:31:01 +00:00
|
|
|
if err := isSentinelValid(sentinel, sp.Policy, sp.Sentinel); err != nil {
|
|
|
|
return nil, fmt.Errorf("Invalid service Sentinel policy: %#v, got error:%v", sp, err)
|
|
|
|
}
|
2014-12-01 03:18:16 +00:00
|
|
|
}
|
|
|
|
|
2016-12-13 04:20:28 +00:00
|
|
|
// Validate the session policies
|
|
|
|
for _, sp := range p.Sessions {
|
|
|
|
if !isPolicyValid(sp.Policy) {
|
|
|
|
return nil, fmt.Errorf("Invalid session policy: %#v", sp)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-06-18 01:56:29 +00:00
|
|
|
// Validate the user event policies
|
|
|
|
for _, ep := range p.Events {
|
Creates new "prepared-query" ACL type and new token capture behavior.
Prior to this change, prepared queries had the following behavior for
ACLs, which will need to change to support templates:
1. A management token, or a token with read access to the service being
queried needed to be provided in order to create a prepared query.
2. The token used to create the prepared query was stored with the query
in the state store and used to execute the query.
3. A management token, or the token used to create the query needed to be
supplied to perform and CRUD operations on an existing prepared query.
This was pretty subtle and complicated behavior, and won't work for
templates since the service name is computed at execution time. To solve
this, we introduce a new "prepared-query" ACL type, where the prefix
applies to the query name for static prepared query types and to the
prefix for template prepared query types.
With this change, the new behavior is:
1. A management token, or a token with "prepared-query" write access to
the query name or (soon) the given template prefix is required to do
any CRUD operations on a prepared query, or to list prepared queries
(the list is filtered by this ACL).
2. You will no longer need a management token to list prepared queries,
but you will only be able to see prepared queries that you have access
to (you get an empty list instead of permission denied).
3. When listing or getting a query, because it was easy to capture
management tokens given the past behavior, this will always blank out
the "Token" field (replacing the contents as <hidden>) for all tokens
unless a management token is supplied. Going forward, we should
discourage people from binding tokens for execution unless strictly
necessary.
4. No token will be captured by default when a prepared query is created.
If the user wishes to supply an execution token then can pass it in via
the "Token" field in the prepared query definition. Otherwise, this
field will default to empty.
5. At execution time, we will use the captured token if it exists with the
prepared query definition, otherwise we will use the token that's passed
in with the request, just like we do for other RPCs (or you can use the
agent's configured token for DNS).
6. Prepared queries with no name (accessible only by ID) will not require
ACLs to create or modify (execution time will depend on the service ACL
configuration). Our argument here is that these are designed to be
ephemeral and the IDs are as good as an ACL. Management tokens will be
able to list all of these.
These changes enable templates, but also enable delegation of authority to
manage the prepared query namespace.
2016-02-23 08:12:58 +00:00
|
|
|
if !isPolicyValid(ep.Policy) {
|
2015-06-18 01:56:29 +00:00
|
|
|
return nil, fmt.Errorf("Invalid event policy: %#v", ep)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
Creates new "prepared-query" ACL type and new token capture behavior.
Prior to this change, prepared queries had the following behavior for
ACLs, which will need to change to support templates:
1. A management token, or a token with read access to the service being
queried needed to be provided in order to create a prepared query.
2. The token used to create the prepared query was stored with the query
in the state store and used to execute the query.
3. A management token, or the token used to create the query needed to be
supplied to perform and CRUD operations on an existing prepared query.
This was pretty subtle and complicated behavior, and won't work for
templates since the service name is computed at execution time. To solve
this, we introduce a new "prepared-query" ACL type, where the prefix
applies to the query name for static prepared query types and to the
prefix for template prepared query types.
With this change, the new behavior is:
1. A management token, or a token with "prepared-query" write access to
the query name or (soon) the given template prefix is required to do
any CRUD operations on a prepared query, or to list prepared queries
(the list is filtered by this ACL).
2. You will no longer need a management token to list prepared queries,
but you will only be able to see prepared queries that you have access
to (you get an empty list instead of permission denied).
3. When listing or getting a query, because it was easy to capture
management tokens given the past behavior, this will always blank out
the "Token" field (replacing the contents as <hidden>) for all tokens
unless a management token is supplied. Going forward, we should
discourage people from binding tokens for execution unless strictly
necessary.
4. No token will be captured by default when a prepared query is created.
If the user wishes to supply an execution token then can pass it in via
the "Token" field in the prepared query definition. Otherwise, this
field will default to empty.
5. At execution time, we will use the captured token if it exists with the
prepared query definition, otherwise we will use the token that's passed
in with the request, just like we do for other RPCs (or you can use the
agent's configured token for DNS).
6. Prepared queries with no name (accessible only by ID) will not require
ACLs to create or modify (execution time will depend on the service ACL
configuration). Our argument here is that these are designed to be
ephemeral and the IDs are as good as an ACL. Management tokens will be
able to list all of these.
These changes enable templates, but also enable delegation of authority to
manage the prepared query namespace.
2016-02-23 08:12:58 +00:00
|
|
|
// Validate the prepared query policies
|
|
|
|
for _, pq := range p.PreparedQueries {
|
|
|
|
if !isPolicyValid(pq.Policy) {
|
2016-02-25 00:57:55 +00:00
|
|
|
return nil, fmt.Errorf("Invalid query policy: %#v", pq)
|
Creates new "prepared-query" ACL type and new token capture behavior.
Prior to this change, prepared queries had the following behavior for
ACLs, which will need to change to support templates:
1. A management token, or a token with read access to the service being
queried needed to be provided in order to create a prepared query.
2. The token used to create the prepared query was stored with the query
in the state store and used to execute the query.
3. A management token, or the token used to create the query needed to be
supplied to perform and CRUD operations on an existing prepared query.
This was pretty subtle and complicated behavior, and won't work for
templates since the service name is computed at execution time. To solve
this, we introduce a new "prepared-query" ACL type, where the prefix
applies to the query name for static prepared query types and to the
prefix for template prepared query types.
With this change, the new behavior is:
1. A management token, or a token with "prepared-query" write access to
the query name or (soon) the given template prefix is required to do
any CRUD operations on a prepared query, or to list prepared queries
(the list is filtered by this ACL).
2. You will no longer need a management token to list prepared queries,
but you will only be able to see prepared queries that you have access
to (you get an empty list instead of permission denied).
3. When listing or getting a query, because it was easy to capture
management tokens given the past behavior, this will always blank out
the "Token" field (replacing the contents as <hidden>) for all tokens
unless a management token is supplied. Going forward, we should
discourage people from binding tokens for execution unless strictly
necessary.
4. No token will be captured by default when a prepared query is created.
If the user wishes to supply an execution token then can pass it in via
the "Token" field in the prepared query definition. Otherwise, this
field will default to empty.
5. At execution time, we will use the captured token if it exists with the
prepared query definition, otherwise we will use the token that's passed
in with the request, just like we do for other RPCs (or you can use the
agent's configured token for DNS).
6. Prepared queries with no name (accessible only by ID) will not require
ACLs to create or modify (execution time will depend on the service ACL
configuration). Our argument here is that these are designed to be
ephemeral and the IDs are as good as an ACL. Management tokens will be
able to list all of these.
These changes enable templates, but also enable delegation of authority to
manage the prepared query namespace.
2016-02-23 08:12:58 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
// Validate the keyring policy - this one is allowed to be empty
|
|
|
|
if p.Keyring != "" && !isPolicyValid(p.Keyring) {
|
2015-07-07 16:45:38 +00:00
|
|
|
return nil, fmt.Errorf("Invalid keyring policy: %#v", p.Keyring)
|
2015-07-07 00:28:09 +00:00
|
|
|
}
|
|
|
|
|
2016-08-30 02:09:57 +00:00
|
|
|
// Validate the operator policy - this one is allowed to be empty
|
|
|
|
if p.Operator != "" && !isPolicyValid(p.Operator) {
|
|
|
|
return nil, fmt.Errorf("Invalid operator policy: %#v", p.Operator)
|
|
|
|
}
|
|
|
|
|
2014-08-06 22:08:17 +00:00
|
|
|
return p, nil
|
|
|
|
}
|