[PATCH] D69753: [InstCombine] Canonicalize ssub.with.overflow to ssub.sat

[Sanjay Patel via Phabricator via llvm-commits]

spatel added a comment.

In D69753#1740120 <https://reviews.llvm.org/D69753#1740120>, @lebedev.ri wrote:

> Such a pattern matching maze, wow.


It can at least be reduced similarly to the sadd_sat patch by bailing out early if Op is not X or Y. And maybe that makes it easier to combine the logic for ssub_sat and sadd_sat, but I'm not sure if that makes the code more confusing than it's worth.

> Looks correct to me, but wouldn't hurt if someone else could also take a look (@nikic? @spatel ?)

Seems like the expected extension of the earlier patch, so LGTM.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D69753/new/

https://reviews.llvm.org/D69753