[all-commits] [llvm/llvm-project] e3b3c9: [InstCombine] Fix missed opportunity to fold 'or' ...

[Craig Topper via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: e3b3c91dd0bbc8bd6f1ee562641daf1e554eb1b6
      https://github.com/llvm/llvm-project/commit/e3b3c91dd0bbc8bd6f1ee562641daf1e554eb1b6
  Author: Craig Topper <craig.topper at sifive.com>
  Date:   2023-12-03 (Sun, 03 Dec 2023)

  Changed paths:
    M llvm/lib/Transforms/InstCombine/InstCombineAndOrXor.cpp
    M llvm/test/Transforms/InstCombine/or.ll

  Log Message:
  -----------
  [InstCombine] Fix missed opportunity to fold 'or' into 'mul' operand. (#74225)

We were able to fold
or (mul X, Y), X --> mul X, (add Y, 1) (when the multiply has no common
bits with X)
    
This patch makes the transform work if the mul operands are commuted.