[llvm] KnownBits: refine high-bits of mul in signed case (PR #113051)

[Ramkumar Ramachandra via llvm-commits]

artagnon wrote:

> Right. I _think_ I can work on a follow-up doing the mul_ov with twice the bitwidth, and make the high bits exhaustive. Will investigate after this patch lands.

I thought about it some more, and have come to the same conclusion as you: making high bits optimal is as hard as making the entire mul optimal. I was hoping to rely on remainders, but urem isn't optimal either.

https://github.com/llvm/llvm-project/pull/113051