[PATCH] D38001: [X86] Don't select any extend of anyext GR32->GR64 to SUBREG_TO_REG. Use INSERT_SUBREG instead.

[Craig Topper via Phabricator via llvm-commits]

craig.topper created this revision.
Herald added a subscriber: wdng.

As far as I know SUBREG_TO_REG is stating that the upper bits are 0. But if we are just converting the GR32 with no checks, then we have no reason to say the upper bits are 0.

Seems to me like we should be selecting this to INSERT_SUBREG.


https://reviews.llvm.org/D38001

Files:
  lib/Target/X86/X86InstrCompiler.td
  test/CodeGen/X86/vector-shuffle-variable-128.ll
  test/CodeGen/X86/vector-shuffle-variable-256.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D38001.115729.patch
Type: text/x-patch
Size: 19809 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20170918/c860d345/attachment.bin>