[libc++][atomic_ref] Use __atomic_fetch_{add,sub} builtins on floating-points whenever possible#135685

dalg24 merged 9 commits intollvm:mainllvm/llvm-project:mainfrom

dalg24:atomic_ref_fetch_add_sub_floating_pointdalg24/llvm-project:atomic_ref_fetch_add_sub_floating_pointCopy head branch name to clipboard