On Thu, Jul 3, 2014 at 3:38 PM, Ilya Enkovich <enkovich....@gmail.com> wrote:
> 2014-07-03 16:07 GMT+04:00 Uros Bizjak <ubiz...@gmail.com>:
>> On Thu, Jul 3, 2014 at 1:50 PM, Ilya Enkovich <enkovich....@gmail.com> wrote:
>>
>>> I didn't find a nice way to fix peephole2 patterns to take register
>>> constraints into account. Is there any way to do it?
>>
>> Use REX_SSE_REGNO_P (REGNO (operands[...])) in the insn C constraint.
>
> Peephole doesn't know whether it works with tuned instruction or not,
> right? I would need to mark all instructions I modify with some
> attribute and then check for it in peephole.
>
>>
>>> Also fully restrict xmm8-15 does not seem right. It is just costly
>>> but not fully disallowed.
>>
>> As said earlier, you can try "Ya*x" as a constraint.
>
> I tried it. It does not seem to affect allocation much. I do not see
> any gain on targeted tests.
Strange, because the documentation claims:
'*'
Says that the following character should be ignored when choosing
register preferences. '*' has no effect on the meaning of the
constraint as a constraint, and no effect on reloading. For LRA
'*' additionally disparages slightly the alternative if the
following character matches the operand.
Let me rethink this a bit. Prehaps we could reconsider Jakub's
proposal with "Ya,!x" (with two alternatives). IIRC this approach was
needed for some MMX alternatives, where we didn't want RA to allocate
a MMX register when the value could be passed in integer regs, but the
value was still allowed in MMX register.
As a side note, I'll investigate pushdf pattern, if it needs to be
updated. The proposed approach is from reload era, and things changed
substantially since then.
Uros.
