[PATCH] D91273: [ARM] Revert low overhead loops with calls before registry allocation.

[Sjoerd Meijer via Phabricator via llvm-commits]

SjoerdMeijer added inline comments.


================
Comment at: llvm/lib/Target/ARM/MVETPAndVPTOptimisationsPass.cpp:162
+bool MVETPAndVPTOptimisations::RevertLoopWithCall(MachineLoop *ML) {
+  LLVM_DEBUG(dbgs() << "MergeLoopEnd on loop " << ML->getHeader()->getName()
+                    << "\n");
----------------
MergeLoopEnd?


================
Comment at: llvm/lib/Target/ARM/MVETPAndVPTOptimisationsPass.cpp:174
+      if (MI.isCall()) {
+        LLVM_DEBUG(dbgs() << "Found call is loop, reverting: " << MI);
+        RevertDoLoopStart(LoopStart, TII);
----------------
Typo: is -> in


================
Comment at: llvm/lib/Target/ARM/MVETPAndVPTOptimisationsPass.cpp:701
+    Modified |= RevertLoopWithCall(ML);
     Modified |= ConvertTailPredLoop(ML, DT);
+  }
----------------
Do we need to run ConvertTailPredLoop if we have reverted above?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D91273/new/

https://reviews.llvm.org/D91273