[llvm] [AMDGPU][gfx1250] Add wait_xcnt before any access that cannot be repeated (PR #168852)

[Jay Foad via llvm-commits]

================
@@ -2059,6 +2066,13 @@ bool SIGfx12CacheControl::enableVolatileAndOrNonTemporal(
   if (IsVolatile) {
     Changed |= setScope(MI, AMDGPU::CPol::SCOPE_SYS);
 
+    if (ST.requiresWaitXCntForSingleAccessInstructions() &&
+        (TII->isFLAT(*MI) || TII->isVBUFFER(*MI))) {
----------------
jayfoad wrote:

Simpler to use isVMEM here and in finalizeStore below?

https://github.com/llvm/llvm-project/pull/168852