[llvm] [AMDGPU][MachineScheduler] Alternative way to control excess RP. (PR #68004)

[Jeffrey Byrnes via llvm-commits]

================
@@ -894,10 +894,22 @@ void GCNSchedStage::setupNewBlock() {
 
 void GCNSchedStage::finalizeGCNRegion() {
   DAG.Regions[RegionIdx] = std::pair(DAG.RegionBegin, DAG.RegionEnd);
-  DAG.RescheduleRegions[RegionIdx] = false;
----------------
jrbyrnes wrote:

This is setting it to false -- the intent seems to be that we don't carry over the flag from previous scheduling stages, and we only set it if RP is still not good after the current stage.

https://github.com/llvm/llvm-project/pull/68004