[PATCH] D117680: [InstCombine] Fold bswap(shl(x, C)) -> and(x, 255)

[Paweł Bylica via Phabricator via llvm-commits]

chfast added a comment.

Hi @spatel,

My motivation is rather boring. I have a generic data -> integer loader. The data is in big-endian order. For i32 example the data can be from 1 to 4 bytes in size. Here is the generic template which handles all 4 cases:
https://godbolt.org/z/dPcGjvb94

The `load<1>` has unnecessary `bswap`, there `zext(*data)` should be enough.
There may be some optimization opportunities in `load<3>` but I did not investigated this in detail.

I can implement other proposed matches on IR and CodeGen levels if you think this is good idea.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D117680/new/

https://reviews.llvm.org/D117680