On Mon, Jan 12, 2026 at 09:22:11AM +0100, Richard Biener wrote:
> I wonder why we assign a mode to op1 in the first place if it is
> VOIDmode and not use widest_int where the CONST_INT to widest_int
> hopefully does the right thing to handle the sign-extended
> representation? (I think we should be able to rely on op1 being
> not negative as well)
The complication is when op1 does have a mode, we should honor that
and so we need wide_int in that case. Sure, we could then convert
it to widest_int but we lack details on if the shift/rotate count
was incorrectly negative or incorrectly too large (or correctly
because SHIFT_COUNT_TRUNCATED and earlier removal of masking).
Jakub
