[PATCH] D48789: [X86] Replace (32/64 - n) shift amounts with (neg n) since the shift amount is masked in hardware

[Craig Topper via Phabricator via llvm-commits]

craig.topper added a comment.

@niravd, let me make sure I understand what will happen in the CSE case. We'll walk the isel matching table using the CSE node, but at the end of the match we'll call MorphNodeTo using the original node that is still held in NodeToMatch in the target independent isel code? Then later we'll revisit the CSE node when its turn in the topological order comes up?


Repository:
  rL LLVM

https://reviews.llvm.org/D48789