[llvm] [AMDGPU] Add scheduling stage to rewrite MFMA from VGPR to AGPR (PR #170335)

[Matt Arsenault via llvm-commits]

================
@@ -1216,6 +1223,107 @@ bool GCNSchedStage::initGCNSchedStage() {
   return true;
 }
 
+void RewriteScheduleStage::findReachingDefs(
+    MachineOperand &UseMO, LiveIntervals *LIS,
+    SmallVectorImpl<SlotIndex> &DefIdxs) {
+  assert(UseMO.isReg());
----------------
arsenm wrote:

```suggestion
```

getReg will just fail the same assert anyway 

https://github.com/llvm/llvm-project/pull/170335