[libc++] std::barrier uses native wait

[huixie90]

not sure why but the current std::barrier's wait implementation is polling on the loop with sleep instead of using the native wait

Before this change, on the CI runner , lost_wakeup.pass.cpp was the slowest test , which uses std::barrier

===
macOS CXX23

Before:

475.32s: llvm-libc++-shared.cfg.in :: std/thread/thread.semaphore/lost_wakeup.pass.cpp

After

10.83s: llvm-libc++-shared.cfg.in :: std/thread/thread.semaphore/lost_wakeup.pass.cpp

===
Linux CXX23:

Before

43.34s: llvm-libc++-shared.cfg.in :: std/thread/thread.semaphore/lost_wakeup.pass.cpp

After:

No longer in the slowest tests list

===
Windows:

Before

80.29s: llvm-libc++-shared-clangcl.cfg.in :: std/thread/thread.semaphore/lost_wakeup.pass.cpp

After

No longer in the slowest test list

This change should also indirectly fix the performance issue of std::barrier described in #123855

[llvmbot]

@llvm/pr-subscribers-libcxx

Author: Hui (huixie90)

Changes

not sure why but the current std::barrier implementation the wait implementation is polling on the loop instead of using the native wait

---

Full diff: https://github.com/llvm/llvm-project/pull/171041.diff

1 Files Affected:

• (modified) libcxx/include/barrier (+1-2)

diff --git a/libcxx/include/barrier b/libcxx/include/barrier
index 5f9b471f01741..10d40f32919c8 100644
--- a/libcxx/include/barrier
+++ b/libcxx/include/barrier
@@ -142,8 +142,7 @@ public:
     return __old_phase;
   }
   _LIBCPP_HIDE_FROM_ABI void wait(arrival_token&& __old_phase) const {
-    auto const __test_fn = [this, __old_phase]() -> bool { return __phase_.load(memory_order_acquire) != __old_phase; };
-    std::__libcpp_thread_poll_with_backoff(__test_fn, __libcpp_timed_backoff_policy());
+    __phase_.wait(__old_phase, std::memory_order_acquire);
   }
   _LIBCPP_HIDE_FROM_ABI void arrive_and_drop() {
     __expected_adjustment_.fetch_sub(1, memory_order_relaxed);