[llvm-branch-commits] [clang] [amdgpu-cfi: 3/9]: [Clang] Default to async unwind tables for amdgcn (PR #183148)

[Matt Arsenault via llvm-branch-commits]

================
@@ -0,0 +1,26 @@
+// REQUIRES: amdgpu-registered-target
+
+// Default options
+// RUN: %clang -### -x cl --target=amdgcn-amd-amdhsa -c -emit-llvm --no-offloadlib -nogpuinc %s 2>&1 | FileCheck -check-prefix=ASYNC-TABLES %s
----------------
arsenm wrote:

though really we should get down to 2. There's pointless divergence between openmp and hip 

https://github.com/llvm/llvm-project/pull/183148