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

Pierre van Houtryve via cfe-commits cfe-commits at lists.llvm.org
Tue Nov 3 23:57:19 PST 2026


================
@@ -3497,6 +3498,43 @@ bool SIInsertWaitcnts::run() {
 
   bool Modified = false;
 
+  if (ST.needsAsyncBarrierArriveReplayWorkaround()) {
----------------
Pierre-vh wrote:

This code doesn't seem to use any of the properties of InsetWaitcnt, why is it here ?
I'd rather avoid adding things to InsertWaitcnt if it can be avoided.

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


More information about the cfe-commits mailing list