[PATCH] D146121: [DAG] Move lshr narrowing from visitANDLike to SimplifyDemandedBits (WIP)

[Simon Pilgrim via Phabricator via llvm-commits]

RKSimon added inline comments.


================
Comment at: llvm/test/CodeGen/X86/bswap.ll:232
+; CHECK64-NEXT:    shlq $8, %rax
+; CHECK64-NEXT:    orq %rcx, %rax
 ; CHECK64-NEXT:    retq
----------------
RKSimon wrote:
> This looks to be the last regression - MatchBSwapHWordLow is very pattern specific, and can't peek through zext (or ignore AND masks for known zero bits)
I've confirmed this is fixed by adding an equivalent SHL narrowing fold in SimplifyDemandedBits, which I intend to do as a follow up.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D146121/new/

https://reviews.llvm.org/D146121