[llvm] [ValueTracking] Respect `samesign` flag in `isKnownInversion` (PR #112390)

[via llvm-commits]

goldsteinn wrote:


Think this definetly makes sense in the case where both operands are non-constants. Its highly unlikely any analysis will add `samesign` to one and not the other.

In the constant RHS case if the constants differ I could image a case where we apply `samesign` to one but not the other. Although I have no issue with this going in as is for now.

https://github.com/llvm/llvm-project/pull/112390