[libc] use LIBC_HAS_BUILTIN instead __has_builtin

[SchrodingerZhu]

fix the issue that __has_builtin is undefined on some non-clang targets.

[llvmbot]

@llvm/pr-subscribers-libc

Author: Schrodinger ZHU Yifan (SchrodingerZhu)

Changes

fix the issue that __has_builtin is undefined on GPU targets.

---

Full diff: https://github.com/llvm/llvm-project/pull/82377.diff

1 Files Affected:

• (modified) libc/src/__support/CPP/atomic.h (+1-1)

diff --git a/libc/src/__support/CPP/atomic.h b/libc/src/__support/CPP/atomic.h
index c10d06c04bccad..b74cb5981dbaf8 100644
--- a/libc/src/__support/CPP/atomic.h
+++ b/libc/src/__support/CPP/atomic.h
@@ -166,7 +166,7 @@ LIBC_INLINE void atomic_thread_fence([[maybe_unused]] MemoryOrder mem_ord) {
 // except no instructions for memory ordering are issued. Only reordering of
 // the instructions by the compiler is suppressed as order instructs.
 LIBC_INLINE void atomic_signal_fence([[maybe_unused]] MemoryOrder mem_ord) {
-#if __has_builtin(__atomic_signal_fence)
+#if LIBC_HAS_BUILTIN(__atomic_signal_fence)
   __atomic_signal_fence(static_cast<int>(mem_ord));
 #else
   // if the builtin is not ready, use asm as a full compiler barrier.

[jhuber6]

The GPU has __has_builtin, that's a clang thing, see https://godbolt.org/z/1MPozEor1. However, we are supposed to be consistent with the LIBC_ prefixed stuff.

[jhuber6]

constexpr if would be a good fit here as well.

[SchrodingerZhu]

it seems that clang would not be happy with unknown function in constexpr if: https://godbolt.org/z/xeGevdqzW

[jhuber6]

Ah, yeah. I remember running into that because constexpr is only evaluated after the AST is built.