[clang] [llvm] [AMDGPU] Add an async barrier arrival workaround (PR #238539)

Yaxun Liu via cfe-commits cfe-commits at lists.llvm.org
Wed Nov 4 11:59:41 PST 2026


yxsamliu wrote:

Moved the workaround to a separate pass just before `SIInsertWaitcnts`, as suggested. It runs after scheduling in both pass managers, including O0.


https://github.com/llvm/llvm-project/pull/238539


More information about the cfe-commits mailing list