Re: [x86 PATCH take #2] Move V1TI shift/rotate lowering from expand to pre-reload split.

On Fri, Aug 12, 2022 at 11:24 PM Roger Sayle wrote: > > > Hi Uros, > As requested, here's an updated version of my patch that introduces a new > const_0_to_255_not_mul_8_operand as you've requested. I think in this > instance, having mutually exclusive patterns that can appear in any order, > wit

[x86 PATCH take #2] Move V1TI shift/rotate lowering from expand to pre-reload split.

Hi Uros, As requested, here's an updated version of my patch that introduces a new const_0_to_255_not_mul_8_operand as you've requested. I think in this instance, having mutually exclusive patterns that can appear in any order, without imposing implicit ordering constraints, is slightly preferabl