[all-commits] [llvm/llvm-project] 949476: [X86] matchShuffleAsBlend - keep per-lane blend ma...

[Simon Pilgrim via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 949476f533838a0d724d184c9f9db856f52fb68d
      https://github.com/llvm/llvm-project/commit/949476f533838a0d724d184c9f9db856f52fb68d
  Author: Simon Pilgrim <llvm-dev at redking.me.uk>
  Date:   2023-01-29 (Sun, 29 Jan 2023)

  Changed paths:
    M llvm/lib/Target/X86/X86ISelLowering.cpp
    M llvm/test/CodeGen/X86/any_extend_vector_inreg_of_broadcast.ll
    M llvm/test/CodeGen/X86/any_extend_vector_inreg_of_broadcast_from_memory.ll
    M llvm/test/CodeGen/X86/avx512-shuffles/partial_permute.ll
    M llvm/test/CodeGen/X86/pr38639.ll
    M llvm/test/CodeGen/X86/pr51615.ll
    M llvm/test/CodeGen/X86/vector-interleaved-load-i16-stride-5.ll
    M llvm/test/CodeGen/X86/vector-interleaved-load-i16-stride-6.ll
    M llvm/test/CodeGen/X86/vector-interleaved-load-i32-stride-3.ll
    M llvm/test/CodeGen/X86/vector-interleaved-load-i32-stride-5.ll
    M llvm/test/CodeGen/X86/vector-interleaved-load-i32-stride-6.ll
    M llvm/test/CodeGen/X86/vector-interleaved-load-i32-stride-7.ll
    M llvm/test/CodeGen/X86/vector-interleaved-load-i64-stride-3.ll
    M llvm/test/CodeGen/X86/vector-interleaved-store-i32-stride-7.ll
    M llvm/test/CodeGen/X86/vector-interleaved-store-i64-stride-5.ll
    M llvm/test/CodeGen/X86/vector-interleaved-store-i64-stride-6.ll
    M llvm/test/CodeGen/X86/vector-interleaved-store-i64-stride-7.ll
    M llvm/test/CodeGen/X86/vector-interleaved-store-i64-stride-8.ll
    M llvm/test/CodeGen/X86/vector-shuffle-256-v4.ll
    M llvm/test/CodeGen/X86/vector-shuffle-256-v8.ll
    M llvm/test/CodeGen/X86/vector-shuffle-combining-avx.ll
    M llvm/test/CodeGen/X86/vector-shuffle-combining-avx2.ll
    M llvm/test/CodeGen/X86/zero_extend_vector_inreg_of_broadcast.ll
    M llvm/test/CodeGen/X86/zero_extend_vector_inreg_of_broadcast_from_memory.ll

  Log Message:
  -----------
  [X86] matchShuffleAsBlend - keep per-lane blend masks 'whole' if only one input is used

For 32/64-bit element 256-bit X86ISD::BLENDI shuffle masks, see if each sub-lane only references (plus undefs) a single input - in which case ensure the blend mask component for that lane ONLY references that input.

This helps later subvector extraction/insertion nodes to peek through the blend node for that lane.