[llvm] [DAG] Prefer sub(X, ~Y) while rewriting inc-of-add to sub-of-not (PR #226379)

[Simon Pilgrim via llvm-commits]

RKSimon wrote:

> No objection from me, I have no reason to see one as better than the other.

Agreed - x86 has a kludge to detect "free" NOTs but I doubt it'd be acceptable to make it generic as a bitwise equivalent to getNegatedExpression.

https://github.com/llvm/llvm-project/pull/226379