[Mlir-commits] [mlir] [MLIR][LLVM] Add `llvm.experimental.constrained.fptrunc` operation (PR #86260)

[Tobias Gysi]

https://github.com/gysit approved this pull request.

Thanks for the changes!

LGTM modulo some minor comments.

Its a shame these intrinsics seem to exists with all kinds of argument number and rounding mode configurations. Otherwise, it may have been nice to factor out base classes unary / binary etc to make things a bit less complex...

https://github.com/llvm/llvm-project/pull/86260