volatile correctness: combine vs. target.md

[Richard Guenther]

On Sat, Dec 3, 2011 at 12:42 AM, Richard Henderson <rth@redhat.com> wrote:
> On 12/02/2011 06:35 AM, Richard Guenther wrote:
>> I see.  As we do not explicitely model this dependency we probably
>> get lucky by the if (gimple_has_volatile_ops ()) bail-out; most
>> passes do.
>
> What are you talking about?  Of course we do.
>
>> int
>> read_dependence (const_rtx mem, const_rtx x)
>> {
>>   return MEM_VOLATILE_P (x) && MEM_VOLATILE_P (mem);
>> }
>
> et al.

Not if you look at the respective gimple level dependency routines
in tree-ssa-alias.c, OTOH tree-data-ref.c simply refuses to handle
volatile references at all.

Richard.

>
> r~