[all-commits] [llvm/llvm-project] e11729: [X86][AVX] canonicalizeLaneShuffleWithRepeatedOps ...

[Simon Pilgrim via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: e1172959226689a7d1b3ba8c04883969161333af
      https://github.com/llvm/llvm-project/commit/e1172959226689a7d1b3ba8c04883969161333af
  Author: Simon Pilgrim <llvm-dev at redking.me.uk>
  Date:   2021-02-06 (Sat, 06 Feb 2021)

  Changed paths:
    M llvm/lib/Target/X86/X86ISelLowering.cpp
    M llvm/test/CodeGen/X86/vector-shuffle-combining-avx.ll

  Log Message:
  -----------
  [X86][AVX] canonicalizeLaneShuffleWithRepeatedOps - merge VPERMILPD ops with different low/high masks.

Now that PR48908 has been dealt with, we can handle v4f64 permute cases by extracting the low/high lane VPERMILPD masks and creating a new mask based on which lanes are referenced by the VPERM2F128 mask.