Support WebAssembly target platforms `wasm{32,64}-unknown-unknown` (#405)

platforms/BUILD.bazel

@@ -45,3 +45,19 @@ platform(
         "@platforms//cpu:aarch64",
     ],
 )
+
+platform(
+    name = "wasm32",
+    constraint_values = [
+        "@platforms//os:none",
+        "@platforms//cpu:wasm32",
+    ],
+)
+
+platform(
+    name = "wasm64",
+    constraint_values = [
+        "@platforms//os:none",
+        "@platforms//cpu:wasm64",
+    ],
+)

tests/MODULE.bazel

@@ -179,3 +179,40 @@ llvm.toolchain(
     exec_arch = "amd64",
 )
 use_repo(llvm, "llvm_toolchain_linux_exec")
+
+# Toolchain example for WebAssembly.
+llvm.toolchain(
+    name = "llvm_toolchain_wasm",
+    # WebAssembly tests use a separate (newer) version of LLVM to exercise
+    # support for experimental features such as wasm64.
+    llvm_versions = {
+        # The most recent LLVM as of 2024-10-17
+        "": "19.1.0",
+    },
+    stdlib = {
+        "wasm32": "libc",
+        "wasm64": "none",
+    },
+    libclang_rt = {
+        "@libclang_rt_wasm32//:libclang_rt.builtins-wasm32.a": "wasm32-unknown-unknown/libclang_rt.builtins.a",
+    },
+)
+llvm.sysroot(
+    name = "llvm_toolchain_wasm",
+    label = "@wasi_sdk_sysroots//wasm32-wasip2",
+    targets = ["wasm32"],
+)
+llvm.sysroot(
+    name = "llvm_toolchain_wasm",
+    label = "@wasi_sdk_sysroots//empty",
+    targets = ["wasm64"],
+)
+
+use_repo(llvm, "llvm_toolchain_wasm")
+register_toolchains("@llvm_toolchain_wasm//:all")
+
+wasi_sdk_sysroots = use_repo_rule("//wasm:wasi_sdk.bzl", "wasi_sdk_sysroots")
+wasi_sdk_sysroots(name = "wasi_sdk_sysroots")
+
+libclang_rt_wasm32 = use_repo_rule("//wasm:wasi_sdk.bzl", "libclang_rt_wasm32")
+libclang_rt_wasm32(name = "libclang_rt_wasm32")

tests/WORKSPACE

@@ -156,6 +156,41 @@ llvm_toolchain(
     urls = {"": ["https://github.com/llvm/llvm-project/releases/download/llvmorg-17.0.6/clang+llvm-17.0.6-x86_64-linux-gnu-ubuntu-22.04.tar.xz"]},
 )
 
+# Toolchain example for WebAssembly.
+llvm_toolchain(
+    name = "llvm_toolchain_wasm",
+    libclang_rt = {
+        "@libclang_rt_wasm32//:libclang_rt.builtins-wasm32.a": "wasm32-unknown-unknown/libclang_rt.builtins.a",
+    },
+    # WebAssembly tests use a separate (newer) version of LLVM to exercise
+    # support for experimental features such as wasm64.
+    llvm_versions = {
+        # The most recent LLVM as of 2024-10-17
+        "": "19.1.0",
+    },
+    stdlib = {
+        "wasm32": "libc",
+        "wasm64": "none",
+    },
+    sysroot = {
+        "wasm32": "@wasi_sdk_sysroots//wasm32-wasip2",
+        "wasm64": "@wasi_sdk_sysroots//empty",
+    },
+)
+
+load(
+    "@llvm_toolchain_wasm//:toolchains.bzl",
+    llvm_register_toolchains_wasm = "llvm_register_toolchains",
+)
+
+llvm_register_toolchains_wasm()
+
+load("//wasm:wasi_sdk.bzl", "libclang_rt_wasm32", "wasi_sdk_sysroots")
+
+libclang_rt_wasm32(name = "libclang_rt_wasm32")
+
+wasi_sdk_sysroots(name = "wasi_sdk_sysroots")
+
 ## Test dependencies.
 
 # Well known repos; present here only for testing.

tests/scripts/debian_test.sh

@@ -30,7 +30,7 @@ set -exuo pipefail
 # Common setup
 export DEBIAN_FRONTEND=noninteractive
 apt-get -qq update
-apt-get -qq -y install curl libtinfo5 zlib1g-dev >/dev/null
+apt-get -qq -y install curl libtinfo5 libxml2 zlib1g-dev >/dev/null
 # The above command gives some verbose output that can not be silenced easily.
 # https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=288778
 

tests/scripts/linux_sysroot_test.sh

@@ -30,7 +30,7 @@ set -exuo pipefail
 # Common setup
 export DEBIAN_FRONTEND=noninteractive
 apt-get -qq update
-apt-get -qq -y install curl libtinfo5 zlib1g-dev >/dev/null
+apt-get -qq -y install curl libtinfo5 libxml2 zlib1g-dev >/dev/null
 # The above command gives some verbose output that can not be silenced easily.
 # https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=288778
 

tests/scripts/run_docker_exec_test.sh

@@ -40,7 +40,7 @@ docker build --platform=linux/amd64 --pull --tag=bazel-docker-sandbox - <<-EOF
 	FROM ${base_image}
 	ENV DEBIAN_FRONTEND=noninteractive
 	RUN apt-get -qq update && \
-		apt-get -qq -y install libtinfo5 zlib1g-dev libxml2
+		apt-get -qq -y install libtinfo5 libxml2 zlib1g-dev libxml2
 EOF
 
 build_args=(

tests/scripts/run_tests.sh

@@ -16,8 +16,9 @@
 set -euo pipefail
 
 toolchain_name=""
+disable_wasm_tests=""
 
-while getopts "t:h" opt; do
+while getopts "t:hW" opt; do
   case "${opt}" in
   "t") toolchain_name="${OPTARG}" ;;
   "h")
@@ -25,6 +26,9 @@ while getopts "t:h" opt; do
     echo "-t - Toolchain name to use for testing; default is llvm_toolchain"
     exit 2
     ;;
+  "W")
+    disable_wasm_tests="yes"
+    ;;
   *)
     echo "invalid option: -${OPTARG}"
     exit 1
@@ -62,3 +66,24 @@ fi
 # Note that the following flags are currently known to cause issues in migration tests:
 # --incompatible_disallow_struct_provider_syntax # https://github.com/bazelbuild/bazel/issues/7347
 # --incompatible_no_rule_outputs_param # from rules_rust
+
+# WebAssembly tests use a separate (newer) version of LLVM to exercise support
+# for experimental features such as wasm64, which can cause the CI environment
+# to run out of disk space.
+#
+# Mitigate this by expunging the workspace before trying to build Wasm targets.
+if [[ -z ${toolchain_name} && -z ${disable_wasm_tests} ]]; then
+  # Redefine `test_args` without `--linkopt=-Wl,-v`, which breaks `wasm-ld`.
+  #
+  # https://github.com/llvm/llvm-project/issues/112836
+  test_args=(
+    "--copt=-v"
+    "--linkopt=-Wl,-t"
+  )
+  wasm_targets=(
+    "//wasm:all"
+  )
+  "${bazel}" clean --expunge
+  "${bazel}" ${TEST_MIGRATION:+"--strict"} --bazelrc=/dev/null test \
+    "${common_test_args[@]}" "${test_args[@]}" "${wasm_targets[@]}"
+fi

tests/scripts/ubuntu_20_04_test.sh

@@ -30,12 +30,16 @@ set -exuo pipefail
 # Common setup
 export DEBIAN_FRONTEND=noninteractive
 apt-get -qq update
-apt-get -qq -y install apt-utils curl libtinfo5 pkg-config zlib1g-dev >/dev/null
+apt-get -qq -y install apt-utils curl libtinfo5 libxml2 pkg-config zlib1g-dev >/dev/null
 # The above command gives some verbose output that can not be silenced easily.
 # https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=288778
 
+# The WebAssembly tests use an LLVM version that is too new for the GNU libc
+# distributed in Ubuntu 20.04.
+disable_wasm_tests='-W'
+
 # Run tests
 cd /src
-tests/scripts/run_tests.sh
+tests/scripts/run_tests.sh \${disable_wasm_tests}
 """
 done

tests/scripts/ubuntu_22_04_test.sh

@@ -30,7 +30,7 @@ set -exuo pipefail
 # Common setup
 export DEBIAN_FRONTEND=noninteractive
 apt-get -qq update
-apt-get -qq -y install curl libtinfo5 zlib1g-dev >/dev/null
+apt-get -qq -y install curl libtinfo5 libxml2 zlib1g-dev >/dev/null
 # The above command gives some verbose output that can not be silenced easily.
 # https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=288778
 

tests/transitions.bzl

@@ -97,11 +97,11 @@ dwp_file = rule(
     },
 )
 
-def _transition_library_to_platform_transition_impl(_, attr):
+def _transition_to_platform_transition_impl(_, attr):
     return {"//command_line_option:platforms": str(attr.platform)}
 
-_transition_library_to_platform_transition = transition(
-    implementation = _transition_library_to_platform_transition_impl,
+_transition_to_platform_transition = transition(
+    implementation = _transition_to_platform_transition_impl,
     inputs = [],
     outputs = ["//command_line_option:platforms"],
 )
@@ -114,7 +114,27 @@ def _transition_library_to_platform_impl(ctx):
 transition_library_to_platform = rule(
     implementation = _transition_library_to_platform_impl,
     attrs = {
-        "lib": attr.label(mandatory = True, cfg = _transition_library_to_platform_transition),
+        "lib": attr.label(mandatory = True, cfg = _transition_to_platform_transition),
+        "platform": attr.label(mandatory = True),
+        "_allowlist_function_transition": attr.label(
+            default = "@bazel_tools//tools/allowlists/function_transition_allowlist",
+        ),
+    },
+)
+
+def _transition_binary_to_platform_impl(ctx):
+    out = ctx.actions.declare_file(ctx.attr.name)
+    ctx.actions.symlink(output = out, target_file = ctx.file.bin)
+    return DefaultInfo(files = depset([out]))
+
+transition_binary_to_platform = rule(
+    implementation = _transition_binary_to_platform_impl,
+    attrs = {
+        "bin": attr.label(
+            mandatory = True,
+            allow_single_file = True,
+            cfg = _transition_to_platform_transition,
+        ),
         "platform": attr.label(mandatory = True),
         "_allowlist_function_transition": attr.label(
             default = "@bazel_tools//tools/allowlists/function_transition_allowlist",

tests/wasm/BUILD.bazel

@@ -0,0 +1,44 @@
+load("@bazel_skylib//rules:build_test.bzl", "build_test")
+load("@rules_cc//cc:defs.bzl", "cc_binary")
+load("//:transitions.bzl", "transition_binary_to_platform")
+
+build_test(
+    name = "wasm_targets_test",
+    targets = [
+        ":wasm32_strlen",
+        ":wasm32_strlen_nolibc",
+        ":wasm64_strlen_nolibc",
+    ],
+)
+
+cc_binary(
+    name = "wasm_strlen",
+    srcs = ["wasm_strlen.c"],
+    linkopts = ["-Wl,--no-entry"],
+    tags = ["manual"],
+)
+
+transition_binary_to_platform(
+    name = "wasm32_strlen",
+    bin = ":wasm_strlen",
+    platform = "@toolchains_llvm//platforms:wasm32",
+)
+
+cc_binary(
+    name = "wasm_strlen_nolibc",
+    srcs = ["wasm_strlen_nolibc.c"],
+    linkopts = ["-Wl,--no-entry"],
+    tags = ["manual"],
+)
+
+transition_binary_to_platform(
+    name = "wasm32_strlen_nolibc",
+    bin = ":wasm_strlen_nolibc",
+    platform = "@toolchains_llvm//platforms:wasm32",
+)
+
+transition_binary_to_platform(
+    name = "wasm64_strlen_nolibc",
+    bin = ":wasm_strlen_nolibc",
+    platform = "@toolchains_llvm//platforms:wasm64",
+)

tests/wasm/wasi_sdk.bzl

@@ -0,0 +1,49 @@
+_SYSROOT_BUILD = """
+filegroup(
+    name = {name},
+    srcs = glob(["include/**/*", "lib/**/*", "share/**/*"], allow_empty=True),
+    visibility = ["//visibility:public"],
+)
+"""
+
+_WASI_SDK_ABIS = [
+    "wasm32-wasi",
+    "wasm32-wasip1",
+    "wasm32-wasip1-threads",
+    "wasm32-wasip2",
+    "wasm32-wasi-threads",
+]
+
+def _wasi_sdk_sysroots(ctx):
+    ctx.download_and_extract(
+        integrity = "sha256-NRcvfSeZSFsVpGsdh/UKWF2RXsZiCA8AXZkVOlCIjwg=",
+        stripPrefix = "wasi-sysroot-24.0",
+        url = ["https://github.com/WebAssembly/wasi-sdk/releases/download/wasi-sdk-24/wasi-sysroot-24.0.tar.gz"],
+    )
+
+    ctx.file("empty/BUILD.bazel", _SYSROOT_BUILD.format(
+        name = repr("empty"),
+    ))
+
+    for abi in _WASI_SDK_ABIS:
+        ctx.file("%s/BUILD.bazel" % (abi,), _SYSROOT_BUILD.format(
+            name = repr(abi),
+        ))
+        ctx.execute(["mv", "include/" + abi, "%s/include" % (abi,)])
+        ctx.execute(["mv", "lib/" + abi, "%s/lib" % (abi,)])
+        ctx.execute(["mv", "share/" + abi, "%s/share" % (abi,)])
+
+wasi_sdk_sysroots = repository_rule(_wasi_sdk_sysroots)
+
+def _libclang_rt_wasm32(ctx):
+    ctx.file("BUILD.bazel", """
+exports_files(glob(["*.a"]))
+""")
+
+    ctx.download_and_extract(
+        integrity = "sha256-fjPA33WLkEabHePKFY4tCn9xk01YhFJbpqNy3gs7Dsc=",
+        stripPrefix = "libclang_rt.builtins-wasm32-wasi-24.0",
+        url = ["https://github.com/WebAssembly/wasi-sdk/releases/download/wasi-sdk-24/libclang_rt.builtins-wasm32-wasi-24.0.tar.gz"],
+    )
+
+libclang_rt_wasm32 = repository_rule(_libclang_rt_wasm32)

tests/wasm/wasm_strlen.c

@@ -0,0 +1,7 @@
+#include <stdint.h>
+#include <string.h>
+
+__attribute__((export_name("strlen")))
+uint32_t wasm_strlen(char *s) {
+	return strlen(s);
+}

tests/wasm/wasm_strlen_nolibc.c

@@ -0,0 +1,6 @@
+__attribute__((export_name("strlen")))
+unsigned long wasm_strlen(char *s) {
+	unsigned long len = 0;
+	for (; *s; len++) {}
+	return len;
+}

toolchain/BUILD.llvm_repo

@@ -41,6 +41,7 @@ filegroup(
     srcs = [
         "bin/ld.lld",
         "bin/ld64.lld",
+        "bin/wasm-ld",
     ],
 )
 

toolchain/cc_toolchain_config.bzl

@@ -89,6 +89,22 @@ def cc_toolchain_config(
             "clang",
             "glibc_unknown",
         ),
+        "wasm32": (
+            "clang-wasm32",
+            "wasm32",
+            "unknown",
+            "clang",
+            "unknown",
+            "unknown",
+        ),
+        "wasm64": (
+            "clang-wasm64",
+            "wasm64",
+            "unknown",
+            "clang",
+            "unknown",
+            "unknown",
+        ),
     }[target_os_arch_key]
 
     # Unfiltered compiler flags; these are placed at the end of the command
@@ -134,10 +150,15 @@ def cc_toolchain_config(
 
     link_flags = [
         "--target=" + target_system_name,
-        "-lm",
         "-no-canonical-prefixes",
     ]
 
+    stdlib = compiler_configuration["stdlib"]
+    if stdlib != "none":
+        link_flags.extend([
+            "-lm",
+        ])
+
     # Similar to link_flags, but placed later in the command line such that
     # unused symbols are not stripped.
     link_libs = []
@@ -166,6 +187,10 @@ def cc_toolchain_config(
         archive_flags.extend([
             "-static",
         ])
+    elif target_arch in ["wasm32", "wasm64"]:
+        # lld is invoked as wasm-ld for WebAssembly targets.
+        use_lld = True
+        use_libtool = False
     else:
         # Note that for xcompiling from darwin to linux, the native ld64 is
         # not an option because it is not a cross-linker, so lld is the
@@ -183,7 +208,6 @@ def cc_toolchain_config(
     # The linker has no way of knowing if there are C++ objects; so we
     # always link C++ libraries.
     cxx_standard = compiler_configuration["cxx_standard"]
-    stdlib = compiler_configuration["stdlib"]
     sysroot_path = compiler_configuration["sysroot_path"]
     if stdlib == "builtin-libc++" and is_xcompile:
         stdlib = "stdc++"
@@ -253,11 +277,14 @@ def cc_toolchain_config(
         link_flags.extend([
             "-l:libstdc++.a",
         ])
+    elif stdlib == "libc":
+        cxx_flags = [
+            "-std=" + cxx_standard,
+        ]
     elif stdlib == "none":
         cxx_flags = [
             "-nostdlib",
         ]
-
         link_flags.extend([
             "-nostdlib",
         ])

toolchain/internal/common.bzl

@@ -12,7 +12,14 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-SUPPORTED_TARGETS = [("linux", "x86_64"), ("linux", "aarch64"), ("darwin", "x86_64"), ("darwin", "aarch64")]
+SUPPORTED_TARGETS = [
+    ("linux", "x86_64"),
+    ("linux", "aarch64"),
+    ("darwin", "x86_64"),
+    ("darwin", "aarch64"),
+    ("none", "wasm32"),
+    ("none", "wasm64"),
+]
 
 # Map of tool name to its symlinked name in the tools directory.
 # See tool_paths in toolchain/cc_toolchain_config.bzl.
@@ -120,7 +127,7 @@ def os(rctx):
 
 def os_bzl(os):
     # Return the OS string as used in bazel platform constraints.
-    return {"darwin": "osx", "linux": "linux"}[os]
+    return {"darwin": "osx", "linux": "linux", "none": "none"}[os]
 
 def arch(rctx):
     arch = rctx.attr.exec_arch
@@ -139,7 +146,12 @@ def arch(rctx):
         return "x86_64"
     return arch
 
+def is_standalone_arch(os, arch):
+    return os == "none" and arch in ["wasm32", "wasm64"]
+
 def os_arch_pair(os, arch):
+    if is_standalone_arch(os, arch):
+        return arch
     return "{}-{}".format(os, arch)
 
 _supported_os_arch = [os_arch_pair(os, arch) for (os, arch) in SUPPORTED_TARGETS]

toolchain/internal/configure.bzl

@@ -25,6 +25,7 @@ load(
     _check_os_arch_keys = "check_os_arch_keys",
     _exec_os_arch_dict_value = "exec_os_arch_dict_value",
     _is_absolute_path = "is_absolute_path",
+    _is_standalone_arch = "is_standalone_arch",
     _list_to_string = "list_to_string",
     _os = "os",
     _os_arch_pair = "os_arch_pair",
@@ -243,6 +244,9 @@ def _cc_toolchains_str(
     cc_toolchains_str = ""
     toolchain_names = []
     for (target_os, target_arch) in _supported_targets:
+        if _is_standalone_arch(target_os, target_arch):
+            suffix = target_arch
+        else:
             suffix = "{}-{}".format(target_arch, target_os)
         cc_toolchain_str = _cc_toolchain_str(
             rctx,
@@ -315,6 +319,8 @@ def _cc_toolchain_str(
         "darwin-aarch64": "aarch64-apple-macosx",
         "linux-aarch64": "aarch64-unknown-linux-gnu",
         "linux-x86_64": "x86_64-unknown-linux-gnu",
+        "wasm32": "wasm32-unknown-unknown",
+        "wasm64": "wasm64-unknown-unknown",
     }[target_pair]
     cxx_builtin_include_directories = [
         toolchain_path_prefix + "include/c++/v1",
@@ -341,6 +347,11 @@ def _cc_toolchain_str(
             _join(sysroot_prefix, "/usr/include"),
             _join(sysroot_prefix, "/System/Library/Frameworks"),
         ])
+    elif target_os == "none":
+        if sysroot_prefix:
+            cxx_builtin_include_directories.extend([
+                _join(sysroot_prefix, "/include"),
+            ])
     else:
         fail("Unreachable")
 

toolchain/internal/llvm_distributions.bzl

@@ -623,6 +623,14 @@ def download_llvm(rctx):
         auth = _get_auth(rctx, urls),
     )
 
+    if rctx.attr.libclang_rt:
+        clang_versions = rctx.path("lib/clang").readdir()
+        for libclang_rt, lib_name in rctx.attr.libclang_rt.items():
+            libclang_rt_content = rctx.read(libclang_rt)
+            for clang_version in clang_versions:
+                lib_path = clang_version.get_child("lib", lib_name)
+                rctx.file(lib_path, libclang_rt_content, legacy_utf8 = False)
+
     updated_attrs = _attr_dict(rctx.attr)
     if update_sha256:
         updated_attrs["sha256"].update([(key, res.sha256)])

toolchain/internal/repo.bzl

@@ -102,6 +102,12 @@ llvm_repo_attrs.update({
               "As with `llvm_mirror`, these sources will take precedence over the official LLVM " +
               "release sources.",
     ),
+    "libclang_rt": attr.label_keyed_string_dict(
+        mandatory = False,
+        doc = ("Additional libclang_rt libraries to overlay into the downloaded LLVM " +
+               "distribution. The key is the label of a libclang_rt library, " +
+               "and the value is `\"{llvm_target_name}/{library_name}.a\"`."),
+    ),
     "netrc": attr.string(
         mandatory = False,
         doc = "Path to the netrc file for authenticated LLVM URL downloads.",