[clang] [Clang][AMDGPU] Stop defaulting to `one-as` for all atomic scopes (PR #120095)

[Matt Arsenault via cfe-commits]

================
@@ -537,7 +537,8 @@ AMDGPUTargetCodeGenInfo::getLLVMSyncScopeID(const LangOptions &LangOpts,
     break;
   }
 
-  if (Ordering != llvm::AtomicOrdering::SequentiallyConsistent) {
+  if (LangOpts.OpenCL &&
----------------
arsenm wrote:

I am deeply opposed to making any __builtin_amdgcn intrinsic behavior language dependent. This is a target builtin with target defined behavior that should behave consistently regardless of language mode. I'd rather just take an OpenCL performance regression than this 

https://github.com/llvm/llvm-project/pull/120095