[PATCH] D30866: [X86] Recognize AVX2 gather instructions during lowering so we can modify the source input when the mask is all ones

[Sanjay Patel via Phabricator via llvm-commits]

spatel accepted this revision.
spatel added a comment.
This revision is now accepted and ready to land.

LGTM. I don't know if the force-to-zero is important either, but since we're generally happy to add xors to avoid partial reg dependencies, that seems fine to me.

With this, there should be no regressions from the clang change for undef in https://reviews.llvm.org/D30834?


https://reviews.llvm.org/D30866