[llvm] [ExpandMemCmp] Optimize ExpandMemCmp to reduce instruction count on x86 (PR #69609)

[Craig Topper via llvm-commits]

topperc wrote:

Wouldn't that byte swap the zeros that were added by the zext into the LSBs? I don't know this code, but that seems weird.

https://github.com/llvm/llvm-project/pull/69609