[llvm] [SelectionDAG] Use Magic Algorithm for Splitting UDIV/UREM by Constant (PR #154968)

[Marius Kamp via llvm-commits]

mskamp wrote:

> There was another patch for solving this problem that was not correctly linked to the issue #146238

Oh, I completely missed that patch. Thank you for the hint.

It seems, though, that the linked patch follows another approach that may not handle some constants (e.g., 67 if I'm not mistaken). Hence, the approaches in the two patches may complement each other. But an integration of both approaches would most likely require measurements and benchmarks to determine when to select which approach to get the best results.

What is the policy on how to proceed with this kind of conflicts?

https://github.com/llvm/llvm-project/pull/154968