[PATCH] D17916: [X86][AVX512] Fixed VPERMT2* shuffle mask decoding and enabled target shuffle combining.

[Simon Pilgrim via llvm-commits]

RKSimon created this revision.
RKSimon added reviewers: delena, igorb, ab.
RKSimon added a subscriber: llvm-commits.
RKSimon set the repository for this revision to rL LLVM.

Patch to add support for target shuffle combining of X86ISD::VPERMV3 nodes, including support for detecting unary shuffles.

This uncovered several issues with the X86ISD::VPERMV shuffle mask decoding of non-64 bit shuffle mask elements - the bit masking wasn't being correctly computed. A similar issue exists with X86ISD::VPERMV nodes that will need to be fixed when this is setup as well, but I will create repro test cases before pushing another patch for review.

Repository:
  rL LLVM

http://reviews.llvm.org/D17916

Files:
  lib/Target/X86/Utils/X86ShuffleDecode.cpp
  lib/Target/X86/X86ISelLowering.cpp
  lib/Target/X86/X86ShuffleDecodeConstantPool.cpp
  test/CodeGen/X86/vector-shuffle-combining-avx512bw.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D17916.49905.patch
Type: text/x-patch
Size: 7857 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20160306/0c48c2bb/attachment.bin>