On 4/16/19 9:35 AM, Michael Matz wrote:
> Hi,
>
> On Tue, 16 Apr 2019, Jeff Law wrote:
>
>>> I still don't think that what you want follows. Imagine this:
>>>
>>> int foo (int i) { ++i; return i; }
>>>
>>> To claim that this function is anything else than const+pure seems weird
>>> (in fact this function doesn't access anything that must lie in memory at
>>> all). Now take your off-the-mill ABI that passes arguments on stack, and
>>> an only slightly bad code generator, i.e. -O0 on i386. You will get an
>>> modification of the argument slot:
>> Ugh. Good point. But then we're back to the problem with handling of
>> REG_EQUIV notes for stores into the argument space. If the callee owns,
>> has been declared pure/const, but can clobber things like you've shown,
>> then a REG_EQUIV note on those slots must be avoided.
>
> Yes, if we want to be extra precise a REG_EQUAL note would be okay before
> the call insn (but not very useful post regalloc), but a REG_EQUIV note is
> plain wrong for argument slots if there are uses of the destination
> register after the call.
Right. We could have a REG_EQUAL note, but I doubt it's helpful at all.
So going back to Jakub's patch... I think the discussion points to
avoiding the REG_EQUIV notes for outgoing argument slots.
Jeff
