[HIP][libclc] Fix race in cmpxchg and change xor to use CAS instead of atomic builtin

[hdelan]

1. Fix a race condition in cmpxchg
2. Change atomic_xor to use a CAS loop instead of atomic builtin. This is needed to merge this in UR [HIP] Revert add prefetch for USM hip allocations a6b8fa66b537753415d24076f… oneapi-src/unified-runtime#936 so that perf regression can be fixed. The long term fix is to use a compiler flag to choose between builtin and safe CAS implementation, but talks upstream may take some time to figure out the details. See [AMDGPU] Add an option to disable unsafe uses of atomic xor llvm/llvm-project#69229

[JackAKirk]

LGTM

[hdelan]

@intel/llvm-gatekeepers we would like to merge this soon if that's OK

[steffenlarsen]

Windows CI build only failed in post-testing cleanup. Merging.

[hdelan]

Thanks @steffenlarsen !