Patchwork [0/4] Updates in support of genericx86*

login
register
mail settings
Submitter Darren Hart
Date Sept. 5, 2013, 12:18 a.m.
Message ID <cover.1378339881.git.dvhart@linux.intel.com>
Download mbox
Permalink /patch/57381/
State New
Headers show

Pull-request

git://git.yoctoproject.org/poky-contrib dvhart/genericx86

Comments

Darren Hart - Sept. 5, 2013, 12:18 a.m.
As this series bridges both oe-core and poky and is best viewed as one series, I
opted to send the series to both lists.

These patches address bugs/feature requests for additional support of
linux-firmware and xorg-video-* drivers for the new meta-yocto-bsp genericx86
and genericx86-64 machines.

Some of the oe-core changes could be contained in the meta-yocto-bsp layer, but
as these changes come from meta-intel, they are needed by multiple layers, and I
felt it might make more sense in oe-core. If people object to the mga or intel
xserver video drivers, I can respin those for meta-yocto-bsp.

The following changes since commit f41b7a7d4d0463a0dfafe6621d01680b81798019:

  bitbake: hob: remove PACKAGE_INSTALL variable setting from hob (2013-09-04 14:18:49 +0100)

are available in the git repository at:

  git://git.yoctoproject.org/poky-contrib dvhart/genericx86
  http://git.yoctoproject.org/cgit.cgi/poky-contrib/log/?h=dvhart/genericx86

Darren Hart (4):
  linux-firmware: Update SRCREV, pull in iwlwifi-7260 support
  xorg-driver-common: Add configure and install appends from meta-intel
  xf86-video-mga: Pull in Matrox MGA support from meta-intel
  genericx86: Create a x86-common.inc base for the x86 BSPs

 meta-yocto-bsp/conf/machine/genericx86-64.conf     |   29 +-----------------
 meta-yocto-bsp/conf/machine/genericx86.conf        |   31 +-------------------
 meta-yocto-bsp/conf/machine/include/x86-common.inc |   26 ++++++++++++++++
 meta/conf/machine/include/ia32-base.inc            |    2 ++
 .../xorg-driver/xf86-video-mga_1.6.2.bb            |   20 +++++++++++++
 .../xorg-driver/xorg-driver-common.inc             |   17 ++++++++++-
 .../linux-firmware/linux-firmware_git.bb           |   10 ++++---
 7 files changed, 72 insertions(+), 63 deletions(-)
 create mode 100644 meta-yocto-bsp/conf/machine/include/x86-common.inc
 create mode 100644 meta/recipes-graphics/xorg-driver/xf86-video-mga_1.6.2.bb
Otavio Salvador - Sept. 5, 2013, 1:16 a.m.
On Wed, Sep 4, 2013 at 9:18 PM, Darren Hart <dvhart@linux.intel.com> wrote:
> As this series bridges both oe-core and poky and is best viewed as one series, I
> opted to send the series to both lists.
>
> These patches address bugs/feature requests for additional support of
> linux-firmware and xorg-video-* drivers for the new meta-yocto-bsp genericx86
> and genericx86-64 machines.
>
> Some of the oe-core changes could be contained in the meta-yocto-bsp layer, but
> as these changes come from meta-intel, they are needed by multiple layers, and I
> felt it might make more sense in oe-core. If people object to the mga or intel
> xserver video drivers, I can respin those for meta-yocto-bsp.

QEMU does not need them so they should be in meta-yocto-bsp.
Darren Hart - Sept. 5, 2013, 3:35 a.m.
On Wed, 2013-09-04 at 22:16 -0300, Otavio Salvador wrote:
> On Wed, Sep 4, 2013 at 9:18 PM, Darren Hart <dvhart@linux.intel.com> wrote:
> > As this series bridges both oe-core and poky and is best viewed as one series, I
> > opted to send the series to both lists.
> >
> > These patches address bugs/feature requests for additional support of
> > linux-firmware and xorg-video-* drivers for the new meta-yocto-bsp genericx86
> > and genericx86-64 machines.
> >
> > Some of the oe-core changes could be contained in the meta-yocto-bsp layer, but
> > as these changes come from meta-intel, they are needed by multiple layers, and I
> > felt it might make more sense in oe-core. If people object to the mga or intel
> > xserver video drivers, I can respin those for meta-yocto-bsp.
> 
> QEMU does not need them so they should be in meta-yocto-bsp.
> 

I'm fine with that if that is the consensus. Does anyone else care to
weigh in?
Ross Burton - Sept. 5, 2013, 9:47 a.m.
On 5 September 2013 04:35, Darren Hart <dvhart@linux.intel.com> wrote:
>> > Some of the oe-core changes could be contained in the meta-yocto-bsp layer, but
>> > as these changes come from meta-intel, they are needed by multiple layers, and I
>> > felt it might make more sense in oe-core. If people object to the mga or intel
>> > xserver video drivers, I can respin those for meta-yocto-bsp.
>>
>> QEMU does not need them so they should be in meta-yocto-bsp.
>
> I'm fine with that if that is the consensus. Does anyone else care to
> weigh in?

By that logic nearly all the X drivers in oe-core should be in
meta-yocto-bsp (synaptics, intel, modesetting, omap, omapfb, keyboard,
mouse), but then we'd be forcing people using anything other than qemu
to use meta-yocto-bsp (or duplicate the recipes).  A more nuanced
rationale for keeping modesettings/intel/etc is that whilst the driver
isn't used on QEMU it is in use on a large number of machines so it
can be centrally maintained in oe-core.  I'm not convinced this holds
for -mga (unless you're reading this from the late 90s) so it should
be in meta-yocto-bsp.

Ross
Richard Purdie - Sept. 5, 2013, 11:06 a.m.
On Thu, 2013-09-05 at 10:47 +0100, Burton, Ross wrote:
> On 5 September 2013 04:35, Darren Hart <dvhart@linux.intel.com> wrote:
> >> > Some of the oe-core changes could be contained in the meta-yocto-bsp layer, but
> >> > as these changes come from meta-intel, they are needed by multiple layers, and I
> >> > felt it might make more sense in oe-core. If people object to the mga or intel
> >> > xserver video drivers, I can respin those for meta-yocto-bsp.
> >>
> >> QEMU does not need them so they should be in meta-yocto-bsp.
> >
> > I'm fine with that if that is the consensus. Does anyone else care to
> > weigh in?
> 
> By that logic nearly all the X drivers in oe-core should be in
> meta-yocto-bsp (synaptics, intel, modesetting, omap, omapfb, keyboard,
> mouse), but then we'd be forcing people using anything other than qemu
> to use meta-yocto-bsp (or duplicate the recipes).  A more nuanced
> rationale for keeping modesettings/intel/etc is that whilst the driver
> isn't used on QEMU it is in use on a large number of machines so it
> can be centrally maintained in oe-core.  I'm not convinced this holds
> for -mga (unless you're reading this from the late 90s) so it should
> be in meta-yocto-bsp.

There is some thinking going on at the moment about merging genericx86
and qemux86 since the two machines are converging and any differences
could be dealt with at runtime. To me, this is a key detail.

I agree that -mga is a push for OE-Core but the other drivers are useful
and I can see a case for moving them towards the core rather than having
several layers needing them.

Cheers,

Richard
Otavio Salvador - Sept. 5, 2013, 11:40 a.m.
On Thu, Sep 5, 2013 at 8:06 AM, Richard Purdie
<richard.purdie@linuxfoundation.org> wrote:
> On Thu, 2013-09-05 at 10:47 +0100, Burton, Ross wrote:
>> On 5 September 2013 04:35, Darren Hart <dvhart@linux.intel.com> wrote:
>> >> > Some of the oe-core changes could be contained in the meta-yocto-bsp layer, but
>> >> > as these changes come from meta-intel, they are needed by multiple layers, and I
>> >> > felt it might make more sense in oe-core. If people object to the mga or intel
>> >> > xserver video drivers, I can respin those for meta-yocto-bsp.
>> >>
>> >> QEMU does not need them so they should be in meta-yocto-bsp.
>> >
>> > I'm fine with that if that is the consensus. Does anyone else care to
>> > weigh in?
>>
>> By that logic nearly all the X drivers in oe-core should be in
>> meta-yocto-bsp (synaptics, intel, modesetting, omap, omapfb, keyboard,
>> mouse), but then we'd be forcing people using anything other than qemu
>> to use meta-yocto-bsp (or duplicate the recipes).  A more nuanced
>> rationale for keeping modesettings/intel/etc is that whilst the driver
>> isn't used on QEMU it is in use on a large number of machines so it
>> can be centrally maintained in oe-core.  I'm not convinced this holds
>> for -mga (unless you're reading this from the late 90s) so it should
>> be in meta-yocto-bsp.
>
> There is some thinking going on at the moment about merging genericx86
> and qemux86 since the two machines are converging and any differences
> could be dealt with at runtime. To me, this is a key detail.
>
> I agree that -mga is a push for OE-Core but the other drivers are useful
> and I can see a case for moving them towards the core rather than having
> several layers needing them.

I am not sure Intel's driver should be in OE-Core; if that's the case
we'd need nouveau and ati drivers there as well as they are largely
used in PCs.