[llvm-branch-commits] [llvm] AMDGPU/GlobalISel: Improve readanylane combines in regbanklegalize (PR #142789)

[Pierre van Houtryve via llvm-branch-commits]

================
@@ -137,7 +138,109 @@ class AMDGPURegBankLegalizeCombiner {
     return {MatchMI, MatchMI->getOperand(1).getReg()};
   }
 
+  std::pair<GUnmerge *, int> tryMatchRALFromUnmerge(Register Src) {
+    MachineInstr *ReadAnyLane = MRI.getVRegDef(Src);
+    if (ReadAnyLane->getOpcode() == AMDGPU::G_AMDGPU_READANYLANE) {
----------------
Pierre-vh wrote:

use early return here?

https://github.com/llvm/llvm-project/pull/142789