[Mlir-commits] [mlir] [mlir] [arith] Fix ceildivsi lowering in arith-expand (PR #133774)

[Jakub Kuderski]

https://github.com/kuhar commented:

I haven't checked the math but looks good overall.

I wonder if we should add some integration tests to exercise some known corner cases -- we do that for wide integer emulation which was also non-trivial to get right: https://github.com/llvm/llvm-project/tree/main/mlir/test/Integration/Dialect/Arith/CPU

https://github.com/llvm/llvm-project/pull/133774