[llvm] [AMDGPU][Scheduler] Scoring system for rematerialization candidates (PR #153092)

[Jeffrey Byrnes via llvm-commits]

================
@@ -2040,6 +2257,13 @@ void GCNScheduleDAGMILive::updateRegionBoundaries(
     RegionBounds.first = NewMI; // Insertion
 }
 
+void GCNScheduleDAGMILive::deleteMI(unsigned RegionIdx, MachineInstr *MI) {
+  updateRegionBoundaries(Regions[RegionIdx], MI, nullptr);
----------------
jrbyrnes wrote:

I think the `updateRegionBoundaries` is broken -- in the case where we are removing the only instruction from the region.

https://github.com/llvm/llvm-project/pull/153092