[PATCH] D146270: [RISCV] Use LBU for extloadi8.

[Craig Topper via Phabricator via llvm-commits]

craig.topper added a comment.

In D146270#4201177 <https://reviews.llvm.org/D146270#4201177>, @jrtc27 wrote:

> Are there any cases like with all the *W optimisations where it would be better to sign-extend, or is i32 special there due to the *W ops and for i8 the only cases where it's beneficial will already have become sextloads?

I think i32 is special. And right now we don't have any optimizations after isel that notice a difference between lbu and lb. For SExtWRemoval they are both viewed as producing a value that has more than 33 sign bits.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D146270/new/

https://reviews.llvm.org/D146270