[MPlayer-dev-eng] [Patch] swscale on gcc4+amd64

Guillaume POIRIER poirierg at gmail.com
Fri Jun 10 17:44:23 CEST 2005


Hi,

On 6/10/05, Ivan Kalvachev <ikalvachev at gmail.com> wrote:
> On 6/8/05, Rich Felker <dalias at aerifal.cx> wrote:
> > On Wed, Jun 08, 2005 at 01:36:40AM +0300, Ivan Kalvachev wrote:
> > > I just want to make an uneducated quess.
> > > In if I understand the patch right, these 2 variables are used as
> > > offsets and are converted to long in order to fit in register.
> > >
> > > How about using constraint "p" ?
> >
> > WTF is constraint "p"? Does it even exist in older gcc versions? I've
> > never seen it.
> 
> info gcc:
> 
> `p'
>      An operand that is a valid memory address is allowed.  This is for
>      "load address" and "push address" instructions.
> 
>      `p' in the constraint must be accompanied by `address_operand' as
>      the predicate in the `match_operand'.  This predicate interprets
>      the mode specified in the `match_operand' as the mode of the memory
>      reference for which the address would be valid.
> -
> 
> It looks like it is invented for RISC, but is also usable on CISC and
> x86 in particular.
> The code generated looks identical as if `r` constraint is used (at
> least on 3.3.5).
> 
> I tried it on gcc 2.95.x and it compiles fine.
> 
> I guess that if it works on gcc-4 we could make an field test.

Great news Ivan. Could you post a patch here so I can make the test ASAP?

Guillaume
-- 
I should have invented the tape recorder,
that way people would have more quotes of
me to put at the bottom of their e-mails
-- Benjamin Franklin




More information about the MPlayer-dev-eng mailing list