[llvm-branch-commits] [llvm] AMDGPU: Handle legal v2bf16 atomicrmw fadd for gfx12 (PR #95930)

[Matt Arsenault via llvm-branch-commits]

================
@@ -1735,8 +1737,11 @@ defm : SIBufferAtomicPat<"SIbuffer_atomic_dec", i64, "BUFFER_ATOMIC_DEC_X2">;
 let OtherPredicates = [HasAtomicCSubNoRtnInsts] in
 defm : SIBufferAtomicPat<"SIbuffer_atomic_csub", i32, "BUFFER_ATOMIC_CSUB", ["noret"]>;
 
-let SubtargetPredicate = isGFX12Plus in {
+let SubtargetPredicate = HasAtomicBufferPkAddBF16Inst in {
   defm : SIBufferAtomicPat_Common<"SIbuffer_atomic_fadd", v2bf16, "BUFFER_ATOMIC_PK_ADD_BF16_VBUFFER">;
----------------
arsenm wrote:

I think these should be inverted, the inner predicate should be the buffer thing 

https://github.com/llvm/llvm-project/pull/95930