[llvm] [InstCombine] Missed optimization: Fold (sext(a) & sext(c1)) == c2 to (a & c1) == c2 (PR #112646)

[Nikita Popov via llvm-commits]

https://github.com/nikic requested changes to this pull request.

The pattern you implemented looks way too specific to me. You are checking for specific constants C1 and C2, but this should be implemented for *any* C1/C2 that satisfy the necessary preconditions from the proof.

https://github.com/llvm/llvm-project/pull/112646