[llvm-branch-commits] [llvm] AMDGPU/GlobalISel: AMDGPURegBankLegalize (PR #112864)

[Nicolai Hähnle via llvm-branch-commits]

================
@@ -217,6 +217,74 @@ bool AMDGPUInstructionSelector::selectCOPY(MachineInstr &I) const {
   return true;
 }
 
+bool AMDGPUInstructionSelector::selectCOPY_SCC_VCC(MachineInstr &I) const {
----------------
nhaehnle wrote:

I don't think it's trivial to avoid this a priori. I agree that a separate cleanup optimization could do it, in any case I'd say it's best left to a separate focused change.

https://github.com/llvm/llvm-project/pull/112864