[PATCH] D37502: [AMDGPU] Fix shouldClusterMemOps to process flat loads

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added inline comments.


================
Comment at: lib/Target/AMDGPU/SIInstrInfo.cpp:368-369
       (isFLAT(FirstLdSt) && isFLAT(SecondLdSt))) {
     FirstDst = getNamedOperand(FirstLdSt, AMDGPU::OpName::vdata);
+    if (!FirstDst)
+      FirstDst = getNamedOperand(FirstLdSt, AMDGPU::OpName::vdst);
----------------
I think falling back like this is more hazardous than checking if it's flat first


https://reviews.llvm.org/D37502