[PATCH] D57516: [PatternMatch] add special-case uaddo matching for increment-by-one

[Sanjay Patel via Phabricator via llvm-commits]

spatel updated this revision to Diff 184526.
spatel added a subscriber: kparzysz.
spatel added a comment.

(Not sure what's going on with Phab updates - but I haven't seen mails for commits corresponding to this review.)

Patch updated:
I forgot to update existing regression tests for this change. 
I was worried that the existing code wasn't restricted enough, and the x86 diffs show that? Ie, using add 1 rather than inc should be considered a regression. I don't know what to make of the Hexagon diff (cc @kparzysz ).

...but there's a twist: I just saw D57520 <https://reviews.llvm.org/D57520> would cause the same set of extra diffs that I neglected to include here originally. What is going on here?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D57516/new/

https://reviews.llvm.org/D57516

Files:
  include/llvm/IR/PatternMatch.h
  test/CodeGen/Hexagon/swp-epilog-phi5.ll
  test/CodeGen/X86/codegen-prepare-uaddo.ll
  test/CodeGen/X86/copy-eflags.ll
  test/CodeGen/X86/peephole-na-phys-copy-folding.ll
  test/CodeGen/X86/rd-mod-wr-eflags.ll
  test/CodeGen/X86/reverse_branches.ll
  test/Transforms/CodeGenPrepare/overflow-intrinsics.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D57516.184526.patch
Type: text/x-patch
Size: 12376 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20190131/c847490a/attachment.bin>