hurd: Fix build with -Werror,-Wswitch

[sthibaul]

We do not handle all architectures, llvm_unreachable is called after the switch, so we can just break.

Fixes https://lab.llvm.org/buildbot/#/builders/19/builds/23702

[llvmbot]

@llvm/pr-subscribers-clang-driver

@llvm/pr-subscribers-clang

Author: Samuel Thibault (sthibaul)

Changes

We do not handle all architectures, llvm_unreachable is called after the switch, so we can just break.

Fixes https://lab.llvm.org/buildbot/#/builders/19/builds/23702

---

Full diff: https://github.com/llvm/llvm-project/pull/78520.diff

1 Files Affected:

• (modified) clang/lib/Driver/ToolChains/Hurd.cpp (+2)

diff --git a/clang/lib/Driver/ToolChains/Hurd.cpp b/clang/lib/Driver/ToolChains/Hurd.cpp
index 5074eda5f41559..36499fb11ad88f 100644
--- a/clang/lib/Driver/ToolChains/Hurd.cpp
+++ b/clang/lib/Driver/ToolChains/Hurd.cpp
@@ -135,6 +135,8 @@ Tool *Hurd::buildAssembler() const {
 
 std::string Hurd::getDynamicLinker(const ArgList &Args) const {
   switch (getArch()) {
+  default:
+    break;
   case llvm::Triple::x86:
     return "/lib/ld.so";
   case llvm::Triple::x86_64:

[fmayer]

Drive-by: is the llvm_unreachable actually unreachable? I.e. we never call this function with this case? Otherwise we are introducing UB in NDEBUG builds

/// In NDEBUG builds, if the platform does not support a builtin unreachable
/// then we call an internal LLVM runtime function. Otherwise the behavior is
/// controlled by the CMake flag
///   -DLLVM_UNREACHABLE_OPTIMIZE
/// * When "ON" (default) llvm_unreachable() becomes an optimizer hint
///   that the current location is not supposed to be reachable: the hint
///   turns such code path into undefined behavior.  On compilers that don't
///   support such hints, prints a reduced message instead and aborts the
///   program.

[sthibaul]

Drive-by: is the llvm_unreachable actually unreachable? I.e. we never call this function with this case?

Hurd triplets are only defined for x86 and x86_64.

[fmayer]

Is there a reason you put this first? Otherwise IMO it is slightly neater to put it last.

[sthibaul]

Is there a reason you put this first?

No particular reason.

Otherwise IMO it is slightly neater to put it last.

Done so.

[DamonFool]

Is there a reason you put this first? Otherwise IMO it is slightly neater to put it last.

This bug had been fixed here:

commit 1b60ddd920e0caadfa85cc7013b559d6453d7e3e
Author: Jie Fu <jiefu@tencent.com>
Date:   Thu Jan 18 07:34:09 2024 +0800

    [Hurd] Fix -Wswitch in Hurd::getDynamicLinker (NFC)
    
    llvm-project/clang/lib/Driver/ToolChains/Hurd.cpp:137:11:
     error: 60 enumeration values not handled in switch: 'UnknownArch', 'arm', 'armeb'... [-Werror,-Wswitch]
      switch (getArch()) {
              ^~~~~~~~~
    1 error generated.

Thanks.