[llvm-branch-commits] [llvm] [AMDGPU] Make amdgpu.noclobber respect acquire operations (PR #219991)

[Krzysztof Drewniak via llvm-branch-commits]

================
@@ -1352,19 +1344,18 @@ define amdgpu_kernel void @one_as_local_barrier(ptr addrspace(1) noalias %out, p
 ; CHECK-NEXT:    fence syncscope("workgroup") release, !mmra [[META2]]
 ; CHECK-NEXT:    tail call void @llvm.amdgcn.s.barrier()
 ; CHECK-NEXT:    fence syncscope("workgroup") acquire, !mmra [[META2]]
-; CHECK-NEXT:    [[LD:%.*]] = load i32, ptr addrspace(1) [[P1]], align 4, !amdgpu.noclobber [[META0]]
+; CHECK-NEXT:    [[LD:%.*]] = load i32, ptr addrspace(1) [[P1]], align 4
----------------
krzysz00 wrote:

Ok, yeah, this is `!noclobber` though, and we should probably fix this PR to keep it that way

https://github.com/llvm/llvm-project/pull/219991