[llvm] [AMDGPU] Mark the async marker intrinsics nocallback (PR #219099)

[Harrison Hao via llvm-commits]

================
@@ -0,0 +1,20 @@
+; RUN: opt -S -mtriple=amdgcn-amd-amdhsa -mcpu=gfx950 -passes=amdgpu-attributor %s | FileCheck %s
----------------
harrisonGPU wrote:

```suggestion
; RUN: opt -S -mtriple=amdgpu9.50 -passes=amdgpu-attributor %s | FileCheck %s
```

https://github.com/llvm/llvm-project/pull/219099