Patchwork mesa-dri: Enable swrast only by default and intel drivers only on IA platform

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
---

+1

Just another thing, I'd prefer to have DRIDRIVERS as ?= so machine can
override it.

On Fri, Oct 14, 2011 at 11:42, Richard Purdie
<richard.purdie@linuxfoundation.org> wrote:
> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
> ---
> diff --git a/meta/recipes-graphics/mesa/mesa-dri.inc b/meta/recipes-graphics/mesa/mesa-dri.inc
> index 603438e..795144a 100644
> --- a/meta/recipes-graphics/mesa/mesa-dri.inc
> +++ b/meta/recipes-graphics/mesa/mesa-dri.inc
> @@ -4,7 +4,11 @@ LIB_DEPS += "libdrm expat"
>  # most of our targets do not have DRI so will use mesa-xlib
>  DEFAULT_PREFERENCE = "-1"
>
> -EXTRA_OECONF += "--with-driver=dri --disable-egl --disable-gallium --without-gallium-drivers"
> +DRIDRIVERS = "swrast"
> +DRIDRIVERS_x86 = "swrast,i915,i965"
> +DRIDRIVERS_x86-64 = "swrast,i915,i965"
> +
> +EXTRA_OECONF += "--with-driver=dri --disable-egl --disable-gallium --without-gallium-drivers --with-dri-drivers=${DRIDRIVERS}"
>
>  python populate_packages_prepend() {
>        import os.path
> diff --git a/meta/recipes-graphics/mesa/mesa-dri_7.11.bb b/meta/recipes-graphics/mesa/mesa-dri_7.11.bb
> index 6a4f9a9..5d25127 100644
> --- a/meta/recipes-graphics/mesa/mesa-dri_7.11.bb
> +++ b/meta/recipes-graphics/mesa/mesa-dri_7.11.bb
> @@ -2,7 +2,3 @@ include mesa-common.inc
>  include mesa-${PV}.inc
>  include mesa-dri.inc
>  PR = "${INC_PR}.0"
> -
> -EXTRA_OECONF += "--with-dri-drivers=swrast,i915,i965"
> -
> -COMPATIBLE_HOST = '(i.86.*-linux|x86_64.*-linux)'
> diff --git a/meta/recipes-graphics/mesa/mesa-dri_git.bb b/meta/recipes-graphics/mesa/mesa-dri_git.bb
> index 15cde60..a0e1945 100644
> --- a/meta/recipes-graphics/mesa/mesa-dri_git.bb
> +++ b/meta/recipes-graphics/mesa/mesa-dri_git.bb
> @@ -7,5 +7,3 @@ DEFAULT_PREFERENCE = "-2"
>
>  PR = "${INC_PR}.0"
>
> -COMPATIBLE_HOST = '(i.86|x86_64).*-linux'
> -
>
>
>
> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core@lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
>

On Fri, 2011-10-14 at 11:46 -0300, Otavio Salvador wrote:
> +1
> 
> Just another thing, I'd prefer to have DRIDRIVERS as ?= so machine can
> override it.

I really wouldn't recommend overriding this on a per machine basis, it
needs to be on a per arch basis. This is because the recipe is not
machine specific (nor should it be).

Configuration therefore falls to the distro, not machine.

Cheers,

Richard

On Fri, Oct 14, 2011 at 04:25:16PM +0100, Richard Purdie wrote:
> On Fri, 2011-10-14 at 11:46 -0300, Otavio Salvador wrote:
> > +1
> > 
> > Just another thing, I'd prefer to have DRIDRIVERS as ?= so machine can
> > override it.
> 
> I really wouldn't recommend overriding this on a per machine basis, it
> needs to be on a per arch basis. This is because the recipe is not
> machine specific (nor should it be).
> 
> Configuration therefore falls to the distro, not machine.

Why not make it machine specific only when machine provides own module
(like the case with glamo on om-gta02)?

Or recipe cannot change PACKAGE_ARCH in some special cases (like
$MACHINE in path to some file in SRC_URI) anymore?

Regards,

Op 14 okt. 2011, om 17:30 heeft Martin Jansa het volgende geschreven:

> On Fri, Oct 14, 2011 at 04:25:16PM +0100, Richard Purdie wrote:
>> On Fri, 2011-10-14 at 11:46 -0300, Otavio Salvador wrote:
>>> +1
>>> 
>>> Just another thing, I'd prefer to have DRIDRIVERS as ?= so machine can
>>> override it.
>> 
>> I really wouldn't recommend overriding this on a per machine basis, it
>> needs to be on a per arch basis. This is because the recipe is not
>> machine specific (nor should it be).
>> 
>> Configuration therefore falls to the distro, not machine.
> 
> Why not make it machine specific only when machine provides own module
> (like the case with glamo on om-gta02)?
> 
> Or recipe cannot change PACKAGE_ARCH in some special cases (like
> $MACHINE in path to some file in SRC_URI) anymore?

Kinda like the wpa-supplicant and mplayer trick in OE classic?


> 
> Regards,
> 
> -- 
> Martin 'JaMa' Jansa     jabber: Martin.Jansa@gmail.com
> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core@lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core

On Fri, 2011-10-14 at 17:30 +0200, Martin Jansa wrote:
> On Fri, Oct 14, 2011 at 04:25:16PM +0100, Richard Purdie wrote:
> > On Fri, 2011-10-14 at 11:46 -0300, Otavio Salvador wrote:
> > > +1
> > > 
> > > Just another thing, I'd prefer to have DRIDRIVERS as ?= so machine can
> > > override it.
> > 
> > I really wouldn't recommend overriding this on a per machine basis, it
> > needs to be on a per arch basis. This is because the recipe is not
> > machine specific (nor should it be).
> > 
> > Configuration therefore falls to the distro, not machine.
> 
> Why not make it machine specific only when machine provides own module
> (like the case with glamo on om-gta02)?
> 
> Or recipe cannot change PACKAGE_ARCH in some special cases (like
> $MACHINE in path to some file in SRC_URI) anymore?

It works just fine but its not nice practise in my opinion for a library
like this and I don't see there is any need in this case. Certainly I
don't see it as something OE-Core should be recommending.

Cheers,

Richard

On Fri, Oct 14, 2011 at 04:55:50PM +0100, Richard Purdie wrote:
> On Fri, 2011-10-14 at 17:30 +0200, Martin Jansa wrote:
> > On Fri, Oct 14, 2011 at 04:25:16PM +0100, Richard Purdie wrote:
> > > On Fri, 2011-10-14 at 11:46 -0300, Otavio Salvador wrote:
> > > > +1
> > > > 
> > > > Just another thing, I'd prefer to have DRIDRIVERS as ?= so machine can
> > > > override it.
> > > 
> > > I really wouldn't recommend overriding this on a per machine basis, it
> > > needs to be on a per arch basis. This is because the recipe is not
> > > machine specific (nor should it be).
> > > 
> > > Configuration therefore falls to the distro, not machine.
> > 
> > Why not make it machine specific only when machine provides own module
> > (like the case with glamo on om-gta02)?
> > 
> > Or recipe cannot change PACKAGE_ARCH in some special cases (like
> > $MACHINE in path to some file in SRC_URI) anymore?
> 
> It works just fine but its not nice practise in my opinion for a library
> like this and I don't see there is any need in this case. Certainly I
> don't see it as something OE-Core should be recommending.

So can I send patches adding my glamo.patch to libdrm and mesa-dri so we
can add glamo to 
DRIDRIVERS_armv4t ?