[PATCH] D68092: [AMDGPU] Invert the handling of skip insertion.

[Nicolai Hähnle via Phabricator via llvm-commits]

nhaehnle added inline comments.


================
Comment at: lib/Target/AMDGPU/SIRemoveShortExecBranches.cpp:112
+
+  TII->analyzeBranch(SrcMBB, TrueMBB, FalseMBB, Cond);
+  if (!FalseMBB)
----------------
analyzeBranch's return value must be checked.


================
Comment at: lib/Target/AMDGPU/SIRemoveShortExecBranches.cpp:116-117
+
+  if (MDT->dominates(TrueMBB, &SrcMBB) ||
+      mustRetainExeczBranch(*FalseMBB, *TrueMBB))
+    return false;
----------------
What's the logic here behind using domination as a criterion?


Repository:
  rL LLVM

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D68092/new/

https://reviews.llvm.org/D68092