[PATCH] D48467: [X86] Recognize an fnma in the presence of an intervening shuffle.

[Sanjay Patel via Phabricator via llvm-commits]

spatel added a comment.

In https://reviews.llvm.org/D48467#1140891, @eraman wrote:

> No, this doesn't deal with this. Should I start from the negate and push it
>  down if it is transitively used by fma?


That seems backwards. Why don't we start from the fma and pattern match an fneg of 1 of its operands (looking through a splat as needed)?

Looks like there's already logic in place for this in X86's combineFMA() with a possibly related FIXME comment?


Repository:
  rL LLVM

https://reviews.llvm.org/D48467