ART: Simplify quasi_atomic.h

Removes fences that duplicate std::atomic_thread_fence().

Bug: 71621075
Test: art/test.py --host -j32
Test: art/test.py --target --64 -j4
Change-Id: I008de4d242d1a3cf4d3f50ce171abbbda647bdaa
diff --git a/runtime/class_linker.cc b/runtime/class_linker.cc
index 412834c..3025818 100644
--- a/runtime/class_linker.cc
+++ b/runtime/class_linker.cc
@@ -6179,7 +6179,7 @@
   // Note that there is a race in the presence of multiple threads and we may leak
   // memory from the LinearAlloc, but that's a tradeoff compared to using
   // atomic operations.
-  QuasiAtomic::ThreadFenceRelease();
+  std::atomic_thread_fence(std::memory_order_release);
   new_conflict_method->SetImtConflictTable(new_table, image_pointer_size_);
   return new_conflict_method;
 }