[clang][driver][wasm] Support -stdlib=libstdc++ for WebAssembly

[tbaeder]

Summary

The WebAssembly toolchain currently supports only -stdlib=libc++
and implicitly assumes the c++ stdlib to be libc++. Change this to also
support libstdc++. The default is still libc++.

[dschuff]

I haven't reviewed this yet, but since we got one of these before and never merged it (https://reviews.llvm.org/D101464) we should probably land one of these.

[sbc100]

clang/lib/Driver/ToolChains/WebAssembly.cpp

434

Where does 11 come from here? Do other drivers hardcode this the same way?

[tbaeder]

clang/lib/Driver/ToolChains/WebAssembly.cpp

434

11 is just the current version. lib/Driver/ToolChains/Gnu.cpp figures out the correct version here via GCCInstallationDetector. I thought hard-coding is find for wasm because the v1 for libc++ above is also hardcoded and Gnu.cpp figures that out dynamically as well (in addLibCxxIncludePaths).

[sbc100]

clang/lib/Driver/ToolChains/WebAssembly.cpp

434

I'd rather not hardcode that current version, that seems rather fragile/wrong.

I agree the v1 is probably also wrong (feel free to add a TODO or update that too).

[tbaeder]

clang/lib/Driver/ToolChains/WebAssembly.cpp

434

Updated both. They both currently fall back to the old v1/11. I'm not sure how the test would otherwise work. As far as I know, the wasm sysroot will basically never contain an actual gcc installation, so GCCInstallationDetector can't be properly used.

[sbc100]

clang/lib/Driver/ToolChains/WebAssembly.cpp

426–439

Can't these 6 lines be removed now? (don't they happen as part of addLibStdCXXIncludePaths?)

484

Looking at Gnu.cpp and Fuscia.cpp it seems like return (or return false) here should be fine.

535

Just early return here if no headers can be found? Looking at Gnu.cpp it seems that addLibStdCxxIncludePaths can simply to nothing if no GCC install is found.

[tbaeder]

clang/lib/Driver/ToolChains/WebAssembly.cpp

426–439

Yep, just a leftover, sorry.

535

I saw that, but I'm not sure if this is correct for wasm. The tests certainly break because they check for the /v1/ (and not /11/) include paths but also use -sysroot=/foo, so the new code doesn't add any flags. Is there a good way to update the tests so they stay functional and useful?

[sbc100]

clang/lib/Driver/ToolChains/WebAssembly.cpp

535

I would take a look at how other platforms test this... perhaps they setup some kind of fake header tree? Or perhaps they don't test these paths at all. For sure they don't depend on the actual system where the tests run, right?

Can you point me to the tests that fail if you simply return empty string like on other platforms?

[tbaeder]

Stop hardcoding v1 or 11 and make the tests work like that.

[tbaeder]

clang/lib/Driver/ToolChains/WebAssembly.cpp

535

The test using /foo as sysroot was just clang/test/Driver/wasm-toolchain.cpp.

The driver tests are shipping a file hierarchy to be used for this, so I switched the tests to use that as well.

[sbc100]

Great! Thanks for working on this. I'm curious what toolchain you are working that uses GNU libstdc++?

[tbaeder]

In D117888#3290628, @sbc100 wrote:

Great! Thanks for working on this. I'm curious what toolchain you are working that uses GNU libstdc++?

I'm looking at libstdc++ enablement in a wasm/emscripten toolchain, so nothing existing/public.

Thanks for the review!

[expnkx]

This patch is not correct.

cqwrteur@Home-Server:~/fast_io_cleanup/fast_io/examples/0001.helloworld$ clang++ -o helloworld helloworld.cc -Ofast -std=c++2b -s -flto -fuse-ld=lld --target=wasm32-wasi --sysroot=$HOME/toolchains/sysroot/wasm32-wasi -fno-exceptions -fno-rtti -fno-unwind-tables -fno-asynchronous-unwind-tables -I../../include -stdlib=libstdc++
In file included from helloworld.cc:1:
In file included from ../../include/fast_io.h:9:
In file included from ../../include/fast_io_hosted.h:17:
In file included from ../../include/fast_io_freestanding.h:12:
In file included from ../../include/fast_io_core.h:11:
In file included from ../../include/fast_io_concept.h:20:
In file included from /home/cqwrteur/toolchains/sysroot/wasm32-wasi/include/c++/12.0.1/concepts:44:
/home/cqwrteur/toolchains/sysroot/wasm32-wasi/include/c++/12.0.1/type_traits:38:10: fatal error: 'bits/c++config.h' file not found
#include <bits/c++config.h>

^~~~~~~~~~~~~~~~~~

1 error generated.

It cannot find bits/c++config.h

wasm32-wasi\include\c++\12.0.1\wasm32-wasi\bits

I have started a revision.
https://reviews.llvm.org/D119056

[tbaeder]

I screwed up using the right Differential Revision link, but this landed.

clang/lib/Driver/ToolChains/WebAssembly.h

void AddClangCXXStdlibIncludeArgs(
       const llvm::opt::ArgList &DriverArgs,
       llvm::opt::ArgStringList &CC1Args) const override;
   void AddCXXStdlibLibArgs(const llvm::opt::ArgList &Args,
                            llvm::opt::ArgStringList &CmdArgs) const override;
   SanitizerMask getSupportedSanitizers() const override;
 
   const char *getDefaultLinker() const override { return "wasm-ld"; }
 
+  CXXStdlibType GetDefaultCXXStdlibType() const override {
+    return ToolChain::CST_Libcxx;
+  }
+
   Tool *buildLinker() const override;
 
   std::string getMultiarchTriple(const Driver &D,
                                  const llvm::Triple &TargetTriple,
                                  StringRef SysRoot) const override;
+
+  void addLibCxxIncludePaths(const llvm::opt::ArgList &DriverArgs,
+                             llvm::opt::ArgStringList &CC1Args) const;
+  void addLibStdCXXIncludePaths(const llvm::opt::ArgList &DriverArgs,
+                                llvm::opt::ArgStringList &CC1Args) const;
 };
 
 } // end namespace toolchains
 } // end namespace driver
 } // end namespace clang
 
 #endif // LLVM_CLANG_LIB_DRIVER_TOOLCHAINS_WEBASSEMBLY_H

clang/lib/Driver/ToolChains/WebAssembly.cpp

 //===--- WebAssembly.cpp - WebAssembly ToolChain Implementation -*- C++ -*-===//
 //
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
 // See https://llvm.org/LICENSE.txt for license information.
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
 //===----------------------------------------------------------------------===//
 
 #include "WebAssembly.h"
 #include "CommonArgs.h"
+#include "Gnu.h"
 #include "clang/Basic/Version.h"
 #include "clang/Config/config.h"
 #include "clang/Driver/Compilation.h"
 #include "clang/Driver/Driver.h"
 #include "clang/Driver/DriverDiagnostic.h"
 #include "clang/Driver/Options.h"
+#include "llvm/Option/ArgList.h"
 #include "llvm/Support/FileSystem.h"
 #include "llvm/Support/Path.h"
-#include "llvm/Option/ArgList.h"
+#include "llvm/Support/VirtualFileSystem.h"
 
 using namespace clang::driver;
 using namespace clang::driver::tools;
 using namespace clang::driver::toolchains;
 using namespace clang;
 using namespace llvm::opt;
 
 /// Following the conventions in https://wiki.debian.org/Multiarch/Tuples,

 ToolChain::RuntimeLibType WebAssembly::GetDefaultRuntimeLibType() const {
   return ToolChain::RLT_CompilerRT;
 }
 
 ToolChain::CXXStdlibType
 WebAssembly::GetCXXStdlibType(const ArgList &Args) const {
   if (Arg *A = Args.getLastArg(options::OPT_stdlib_EQ)) {
     StringRef Value = A->getValue();
-    if (Value != "libc++")
+    if (Value == "libc++")
+      return ToolChain::CST_Libcxx;
+    else if (Value == "libstdc++")
+      return ToolChain::CST_Libstdcxx;
+    else
       getDriver().Diag(diag::err_drv_invalid_stdlib_name)
           << A->getAsString(Args);
   }
   return ToolChain::CST_Libcxx;
 }
 
 void WebAssembly::AddClangSystemIncludeArgs(const ArgList &DriverArgs,
                                             ArgStringList &CC1Args) const {
const std::string MultiarchTriple =
         getMultiarchTriple(D, getTriple(), D.SysRoot);
     addSystemInclude(DriverArgs, CC1Args, D.SysRoot + "/include/" + MultiarchTriple);
   }
   addSystemInclude(DriverArgs, CC1Args, D.SysRoot + "/include");
 }
 
 void WebAssembly::AddClangCXXStdlibIncludeArgs(const ArgList &DriverArgs,
                                                ArgStringList &CC1Args) const {
-  if (!DriverArgs.hasArg(options::OPT_nostdlibinc) &&
-      !DriverArgs.hasArg(options::OPT_nostdincxx)) {
-    if (getTriple().getOS() != llvm::Triple::UnknownOS) {
-      const std::string MultiarchTriple =
-          getMultiarchTriple(getDriver(), getTriple(), getDriver().SysRoot);
-      addSystemInclude(DriverArgs, CC1Args,
-                       getDriver().SysRoot + "/include/" + MultiarchTriple +
-                           "/c++/v1");
-    }
-    addSystemInclude(DriverArgs, CC1Args,
-                     getDriver().SysRoot + "/include/c++/v1");
+
+  if (DriverArgs.hasArg(options::OPT_nostdlibinc) ||
+      DriverArgs.hasArg(options::OPT_nostdincxx))
+    return;
+
+  switch (GetCXXStdlibType(DriverArgs)) {
+  case ToolChain::CST_Libcxx:
+    addLibCxxIncludePaths(DriverArgs, CC1Args);
+    break;
+  case ToolChain::CST_Libstdcxx:
+    addLibStdCXXIncludePaths(DriverArgs, CC1Args);
+    break;
+  default:
+    llvm_unreachable("Unknown stdlib type");
   }
 }
 
 void WebAssembly::AddCXXStdlibLibArgs(const llvm::opt::ArgList &Args,
                                       llvm::opt::ArgStringList &CmdArgs) const {
 
   switch (GetCXXStdlibType(Args)) {
   case ToolChain::CST_Libcxx:
     CmdArgs.push_back("-lc++");
     CmdArgs.push_back("-lc++abi");
     break;
   case ToolChain::CST_Libstdcxx:
+    CmdArgs.push_back("-lstdc++");
+    break;
+  default:
     llvm_unreachable("invalid stdlib name");
   }
 }
 
 SanitizerMask WebAssembly::getSupportedSanitizers() const {
   SanitizerMask Res = ToolChain::getSupportedSanitizers();
   if (getTriple().isOSEmscripten()) {
     Res |= SanitizerKind::Vptr | SanitizerKind::Leak | SanitizerKind::Address;
   }
   return Res;
 }
 
 Tool *WebAssembly::buildLinker() const {
   return new tools::wasm::Linker(*this);
 }
+
+void WebAssembly::addLibCxxIncludePaths(
+    const llvm::opt::ArgList &DriverArgs,
+    llvm::opt::ArgStringList &CC1Args) const {
+  const Driver &D = getDriver();
+  std::string SysRoot = computeSysRoot();
+  std::string LibPath = SysRoot + "/include";
+  const std::string MultiarchTriple =
+      getMultiarchTriple(D, getTriple(), SysRoot);
+  bool IsKnownOs = (getTriple().getOS() != llvm::Triple::UnknownOS);
+
+  std::string Version = detectLibcxxVersion(LibPath);
+  if (Version.empty())
+    return;
+
+  // First add the per-target include path if the OS is known.
+  if (IsKnownOs) {
+    std::string TargetDir = LibPath + "/" + MultiarchTriple + "/c++/" + Version;
+    addSystemInclude(DriverArgs, CC1Args, TargetDir);
+  }
+
+  // Second add the generic one.
+  addSystemInclude(DriverArgs, CC1Args, LibPath + "/c++/" + Version);
+}
+
+void WebAssembly::addLibStdCXXIncludePaths(
+    const llvm::opt::ArgList &DriverArgs,
+    llvm::opt::ArgStringList &CC1Args) const {
+  // We cannot use GCCInstallationDetector here as the sysroot usually does
+  // not contain a full GCC installation.
+  // Instead, we search the given sysroot for /usr/include/xx, similar
+  // to how we do it for libc++.
+  const Driver &D = getDriver();
+  std::string SysRoot = computeSysRoot();
+  std::string LibPath = SysRoot + "/include";
+  const std::string MultiarchTriple =
+      getMultiarchTriple(D, getTriple(), SysRoot);
+  bool IsKnownOs = (getTriple().getOS() != llvm::Triple::UnknownOS);
+
+  // This is similar to detectLibcxxVersion()
+  std::string Version = "";
+  {
+    std::error_code EC;
+    Generic_GCC::GCCVersion MaxVersion =
+        Generic_GCC::GCCVersion::Parse("0.0.0");
+    SmallString<128> Path(LibPath);
+    llvm::sys::path::append(Path, "c++");
+    for (llvm::vfs::directory_iterator LI = getVFS().dir_begin(Path, EC), LE;
+         !EC && LI != LE; LI = LI.increment(EC)) {
+      StringRef VersionText = llvm::sys::path::filename(LI->path());
+      if (VersionText[0] != 'v') {
+        auto Version = Generic_GCC::GCCVersion::Parse(VersionText);
+        if (Version > MaxVersion)
+          MaxVersion = Version;
+      }
+    }
+    if (MaxVersion.Major > 0)
+      Version = MaxVersion.Text;
+  }
+
+  if (Version.empty())
+    return;
+
+  // First add the per-target include path if the OS is known.
+  if (IsKnownOs) {
+    std::string TargetDir = LibPath + "/" + MultiarchTriple + "/c++/" + Version;
+    addSystemInclude(DriverArgs, CC1Args, TargetDir);
+  }
+
+  // Second add the generic one.
+  addSystemInclude(DriverArgs, CC1Args, LibPath + "/c++/" + Version);
+}

clang/test/Driver/wasm-toolchain.cpp

 
 // A basic C++ link command-line with unknown OS.
 
 // RUN: %clangxx -### -no-canonical-prefixes -target wasm32-unknown-unknown --sysroot=/foo --stdlib=libc++ %s 2>&1 \
 // RUN:   | FileCheck -check-prefix=LINK %s
 // LINK: clang{{.*}}" "-cc1" {{.*}} "-o" "[[temp:[^"]*]]"
 // LINK: wasm-ld{{.*}}" "-L/foo/lib" "crt1.o" "[[temp]]" "-lc++" "-lc++abi" "-lc" "{{.*[/\\]}}libclang_rt.builtins-wasm32.a" "-o" "a.out"
 
+// RUN: %clangxx -### -no-canonical-prefixes -target wasm32-unknown-unknown --sysroot=/foo --stdlib=libstdc++ %s 2>&1 \
+// RUN:   | FileCheck -check-prefix=LINK_STDCXX %s
+// LINK_STDCXX: clang{{.*}}" "-cc1" {{.*}} "-o" "[[temp:[^"]*]]"
+// LINK_STDCXX: wasm-ld{{.*}}" "-L/foo/lib" "crt1.o" "[[temp]]" "-lstdc++" "-lc" "{{.*[/\\]}}libclang_rt.builtins-wasm32.a" "-o" "a.out"
+
 // A basic C++ link command-line with optimization with unknown OS.
 
 // RUN: %clangxx -### -O2 -no-canonical-prefixes -target wasm32-unknown-unknown --sysroot=/foo %s --stdlib=libc++ 2>&1 \
 // RUN:   | FileCheck -check-prefix=LINK_OPT %s
 // LINK_OPT: clang{{.*}}" "-cc1" {{.*}} "-o" "[[temp:[^"]*]]"
 // LINK_OPT: wasm-ld{{.*}}" "-L/foo/lib" "crt1.o" "[[temp]]" "-lc++" "-lc++abi" "-lc" "{{.*[/\\]}}libclang_rt.builtins-wasm32.a" "-o" "a.out"
 
+// RUN: %clangxx -### -O2 -no-canonical-prefixes -target wasm32-unknown-unknown --sysroot=/foo %s --stdlib=libstdc++ 2>&1 \
+// RUN:   | FileCheck -check-prefix=LINK_OPT_STDCXX %s
+// LINK_OPT_STDCXX: clang{{.*}}" "-cc1" {{.*}} "-o" "[[temp:[^"]*]]"
+// LINK_OPT_STDCXX: wasm-ld{{.*}}" "-L/foo/lib" "crt1.o" "[[temp]]" "-lstdc++" "-lc" "{{.*[/\\]}}libclang_rt.builtins-wasm32.a" "-o" "a.out"
+
 // A basic C++ link command-line with known OS.
 
 // RUN: %clangxx -### -no-canonical-prefixes -target wasm32-wasi --sysroot=/foo --stdlib=libc++ %s 2>&1 \
 // RUN:   | FileCheck -check-prefix=LINK_KNOWN %s
 // LINK_KNOWN: clang{{.*}}" "-cc1" {{.*}} "-o" "[[temp:[^"]*]]"
 // LINK_KNOWN: wasm-ld{{.*}}" "-L/foo/lib/wasm32-wasi" "crt1.o" "[[temp]]" "-lc++" "-lc++abi" "-lc" "{{.*[/\\]}}libclang_rt.builtins-wasm32.a" "-o" "a.out"
 
+// RUN: %clangxx -### -no-canonical-prefixes -target wasm32-wasi --sysroot=/foo --stdlib=libstdc++ %s 2>&1 \
+// RUN:   | FileCheck -check-prefix=LINK_KNOWN_STDCXX %s
+// LINK_KNOWN_STDCXX: clang{{.*}}" "-cc1" {{.*}} "-o" "[[temp:[^"]*]]"
+// LINK_KNOWN_STDCXX: wasm-ld{{.*}}" "-L/foo/lib/wasm32-wasi" "crt1.o" "[[temp]]" "-lstdc++" "-lc" "{{.*[/\\]}}libclang_rt.builtins-wasm32.a" "-o" "a.out"
+
 // A basic C++ link command-line with optimization with known OS.
 
 // RUN: %clangxx -### -O2 -no-canonical-prefixes -target wasm32-wasi --sysroot=/foo %s --stdlib=libc++ 2>&1 \
 // RUN:   | FileCheck -check-prefix=LINK_OPT_KNOWN %s
 // LINK_OPT_KNOWN: clang{{.*}}" "-cc1" {{.*}} "-o" "[[temp:[^"]*]]"
 // LINK_OPT_KNOWN: wasm-ld{{.*}}" "-L/foo/lib/wasm32-wasi" "crt1.o" "[[temp]]" "-lc++" "-lc++abi" "-lc" "{{.*[/\\]}}libclang_rt.builtins-wasm32.a" "-o" "a.out"
 
+// RUN: %clangxx -### -O2 -no-canonical-prefixes -target wasm32-wasi --sysroot=/foo %s --stdlib=libstdc++ 2>&1 \
+// RUN:   | FileCheck -check-prefix=LINK_OPT_KNOWN_STDCXX %s
+// LINK_OPT_KNOWN_STDCXX: clang{{.*}}" "-cc1" {{.*}} "-o" "[[temp:[^"]*]]"
+// LINK_OPT_KNOWN_STDCXX: wasm-ld{{.*}}" "-L/foo/lib/wasm32-wasi" "crt1.o" "[[temp]]" "-lstdc++" "-lc" "{{.*[/\\]}}libclang_rt.builtins-wasm32.a" "-o" "a.out"
+
 // A basic C++ compile command-line with known OS.
 
-// RUN: %clangxx -### -no-canonical-prefixes -target wasm32-wasi --sysroot=/foo --stdlib=libc++ %s 2>&1 \
+// RUN: %clangxx -### -no-canonical-prefixes -target wasm32-wasi --stdlib=libc++ %s 2>&1 \
+// RUN:     --sysroot=%S/Inputs/basic_linux_libcxx_tree/usr \
 // RUN:   | FileCheck -check-prefix=COMPILE %s
 // COMPILE: clang{{.*}}" "-cc1"
 // COMPILE: "-resource-dir" "[[RESOURCE_DIR:[^"]*]]"
-// COMPILE: "-isysroot" "/foo"
-// COMPILE: "-internal-isystem" "/foo/include/wasm32-wasi/c++/v1"
-// COMPILE: "-internal-isystem" "/foo/include/c++/v1"
+// COMPILE: "-isysroot" "[[SYSROOT:[^"]+]]"
+// COMPILE: "-internal-isystem" "[[SYSROOT:[^"]+]]/include/wasm32-wasi/c++/v1"
+// COMPILE: "-internal-isystem" "[[SYSROOT:[^"]+]]/include/c++/v1"
 // COMPILE: "-internal-isystem" "[[RESOURCE_DIR]]{{(/|\\\\)}}include"
-// COMPILE: "-internal-isystem" "/foo/include/wasm32-wasi"
-// COMPILE: "-internal-isystem" "/foo/include"
+// COMPILE: "-internal-isystem" "[[SYSROOT:[^"]+]]/include/wasm32-wasi"
+// COMPILE: "-internal-isystem" "[[SYSROOT:[^"]+]]/include"
+
+// RUN: %clangxx -### -no-canonical-prefixes -target wasm32-wasi --stdlib=libstdc++ %s 2>&1 \
+// RUN:     --sysroot=%S/Inputs/basic_linux_libstdcxx_libcxxv2_tree/usr \
+// RUN:   | FileCheck -check-prefix=COMPILE_STDCXX %s
+// COMPILE_STDCXX: clang{{.*}}" "-cc1"
+// COMPILE_STDCXX: "-resource-dir" "[[RESOURCE_DIR:[^"]*]]"
+// COMPILE_STDCXX: "-isysroot" "[[SYSROOT:[^"]+]]"
+// COMPILE_STDCXX: "-internal-isystem" "[[SYSROOT:[^"]+]]/include/wasm32-wasi/c++/4.8"
+// COMPILE_STDCXX: "-internal-isystem" "[[SYSROOT:[^"]+]]/include/c++/4.8"
+// COMPILE_STDCXX: "-internal-isystem" "[[RESOURCE_DIR]]{{(/|\\\\)}}include"
+// COMPILE_STDCXX: "-internal-isystem" "[[SYSROOT:[^"]+]]/include/wasm32-wasi"
+// COMPILE_STDCXX: "-internal-isystem" "[[SYSROOT:[^"]+]]/include"