[PATCH] D98614: [AMDGPU] Fix shortfalls in WQM marking

[Piotr Sobczak via Phabricator via llvm-commits]

piotr added inline comments.


================
Comment at: llvm/lib/Target/AMDGPU/SIWholeQuadMode.cpp:365
+      // Find next predecessor to process
+      unsigned Idx = 0;
+      auto PI = MBB->pred_begin() + NextPredIdx;
----------------
To keep the original behaviour, you also need to adjust Idx here and assign it to NextPredIdx, because Idx is used after the for loop in line 377, right?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D98614/new/

https://reviews.llvm.org/D98614