[PATCH] D125232: [riscv] Fix data flow for VSETVLI insertion

Fraser Cormack via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Mon May 9 07:53:05 PDT 2022


frasercrmck added a comment.

Looks decent to me. It overlaps with what I have in D125021 <https://reviews.llvm.org/D125021> and was playing around with locally, but improves on it nicely. Would be good to hear @craig.topper's and @rogfer01's thoughts if possible.



================
Comment at: llvm/lib/Target/RISCV/RISCVInsertVSETVLI.cpp:387
-    // and should keep the predecessor.
-    if (isCompatible(Other, /*Strict*/ true))
-      return *this;
----------------
This is the last use of `/*Strict*/ true` - can we follow up on this and get rid of `Strict` entirely, which was always a patch job?


================
Comment at: llvm/lib/Target/RISCV/RISCVInsertVSETVLI.cpp:970
 void RISCVInsertVSETVLI::computeIncomingVLVTYPE(const MachineBasicBlock &MBB) {
+
   BlockData &BBInfo = BlockInfo[MBB.getNumber()];
----------------
Unintended new line?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D125232/new/

https://reviews.llvm.org/D125232



More information about the llvm-commits mailing list