[clang] [llvm] [AMDGPU] Introduce asyncmark/wait intrinsics (PR #173259)

[Ryan Mitchell via cfe-commits]

================
@@ -368,12 +368,15 @@ enum CPol {
   GLC = 1,
   SLC = 2,
   DLC = 4,
+  SWZ_pregfx12 = 8,
   SCC = 16,
+  ASYNC_pregfx12 = 32,
----------------
RyanRio wrote:

I think in this case even if autoupgrade is correct to use it makes more sense to use separate async versions of the intrinsic

https://github.com/llvm/llvm-project/pull/173259