[MPlayer-dev-eng] [PATCH] Athlon 64 configure script support

[Guillaume POIRIER]

Hi,

On 5/26/05, Corey Hickey <bugfood-ml at fatooh.org> wrote:
> Guillaume POIRIER wrote:
> > Hi,
> 
> > The patch seems fine.
> > Tested in 32-bit mode with gcc-2.95 and 3.4.4. Works fine. I'll try it
> > in 64 bit mode ASAP.
> >
> > Guillaume
> >
> 
> Unless I'm mistaken, the patch doesn't affect 64-bit operation because
> it lies within the:
> 
> case "$host_arch" in
>   i[3-9]86|x86|x86pc|k5|k6|k6-2|k6-3|pentium*|athlon*|i586-i686)
> 
> 
> I've attached a patch for the amd64|x86_64 case. It might or might not
> be applicable as-is; this is my first time really delving into the
> configure file and I'm not familiar with it at all. I've tested my patch
> with gcc versions 3.3.6 and 3.4.4; mplayer configures, builds, and runs
> on each.

Works with the same version of gcc here, plus gcc-4.0 and gcc-snapshot
in i386 and AMD-64.


> Some notes:
> 1. If runtime cpu detection is enabled, _mcpu and _march are both set to
> the empty string. I don't know if this is the right thing to do; the
> original version sets both of them empty regardless.

Seems fine to me.


> 2. Unless I'm mistaken, this patch will prevent building mplayer on a
> 64-bit amd64 system with a gcc earlier than 3.3. Guillaume mentioned
> that "In 64-bit mode, that's not much of a problem as I think only
> gcc-3.3, 3.4 and up are officially supported,". I, as well, think that
> is the case, but I haven't been able to verify it. Is the consensus that
> this is true?

I double-checked: my assumption seems correct: all binary
distributions ship nothing previous to gcc-3.3... so someone who would
have some exotic version of gcc should know what he's doing and be
able to fix it himself [or herself :-)].
In addition, recent version of gcc improved quite a bit the quality of
the code generation, so it's not worth the trouble to support those
special cases.


> 3. The original configure doesn't include support for --target on x86_64
> arch. Neither does my patch. Should support for --target be there?

Don't know.


I plan to commit this patch friday around noon (GMT). If you see
obsious problems with it, please speak up so we can discuss it.
Otherwise, it could still be fixed later on.

Regards,
Guillaume