[Mlir-commits] [mlir] [MLIR] Unconditionally take min of max lhs/rhs value in inferRemU (PR #110169)

[llvmlistbot at llvm.org]

goldsteinn wrote:

> I think the idea of "remu can't be larger than either operand unconditionally" makes sense in principle - but it seems like in this case we are also getting rid of the undefined behavior in case the RHS is 0?

Well its undefined, so we can interpret it as we like (and assume it doesn't happen). Thats atleast what we do in llvm ir.

https://github.com/llvm/llvm-project/pull/110169