[Mlir-commits] [mlir] [mlir][arith] Generalize ceil and floor division expansion (PR #206249)

[llvmlistbot at llvm.org]

================

----------------
mygitljf wrote:

I changed the implementation to avoid this issue. The ceil/floor expansions are now pure Arith: I compute one quotient, derive an `i1` adjustment, convert it to the result type, and add or subtract it. Nothing materializes Tensor operations or requires a tensor lowering pipeline anymore.

https://github.com/llvm/llvm-project/pull/206249