[llvm-branch-commits] [clang] [libclc] clang/AMDGPU: Do not look for rocm device libs if environment is llvm (PR #180922)

[Joseph Huber via llvm-branch-commits]

================
@@ -154,7 +154,7 @@ endif()
 # List of all targets. Note that some are added dynamically below.
 set( LIBCLC_TARGETS_ALL
   amdgcn--
-  amdgcn-amd-amdhsa
+  amdgcn-amd-amdhsa-llvm
----------------
jhuber6 wrote:

This bit will make passing `amdgcn-amd-amdhsa` fail now. I wish the CMake logic here wasn't so restrictive. Why is this needed for this PR right now? I think we pass `-nostdlib` which might already suppress the case in this PR.

https://github.com/llvm/llvm-project/pull/180922