[PATCH] D128700: [AMDGPU][Clang] Skip adding noundef attribute to AMDGPU HIP device functions

[krishna chaitanya sankisa via Phabricator via cfe-commits]

skc7 added inline comments.


================
Comment at: clang/lib/CodeGen/CGCall.cpp:2309
+  bool EnableNoundefAttrs = CodeGenOpts.EnableNoundefAttrs &&
+                            !(getLangOpts().HIP && getLangOpts().CUDAIsDevice);
+
----------------
arsenm wrote:
> Shouldn't be limited to HIP. All languages with convergent operations
With this above suggested change to include all languages with convergent operations, lot of tests had to be updated. Created this new patch D128907 for same.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D128700/new/

https://reviews.llvm.org/D128700