Re: Bug#556213: please do upgrade to xserver-xorg-video-intel 2.9.1
[moving away from the bug since this gets off-topic]
On Wed, Nov 18, 2009 at 17:55:54 -0500, Daniel Kahn Gillmor wrote:
> On 11/18/2009 04:43 PM, Julien Cristau wrote:
> > Having maintainers for X in Debian would be a Nice Thing ;)
>
> heh. Touché. And thanks for all your work so far, Julien!
>
> > 2.9.1 is mostly ready in git, it needs someone to prepare an upload.
> > If you want to help I'm happy to add you to the pkg-xorg alioth group :)
>
> i've never read the source for the intel drivers, and i really don't
> understand graphics hardware, i'm afraid. I just built the 2.9.1 source
> because it was so easy to build by applying the 2.9.0 diff.gz (and
> because upstream seems to want bug reports to come from folks running
> bleeding-edge stuff [0])
>
> However, looking more at debian's diff.gz, i'm particularly unclear
> about the purpose and meaning of the .g4i files added to
> src/xvmc/shader/{mc,vld}. if those are the preferred form of
> modification, i don't think i'm really qualified to sign off on them,
> let alone to be able to judge the desirability of upstream modifications
> for debian :(
>
They're assembly for the gen4 hardware (965 and up). I have to admit
I've never looked at that language. It's probably documented in the
i965 docs, and the assembler is at
http://cgit.freedesktop.org/xorg/app/intel-gen4asm/
> > (One thing to consider is whether to enable kms with the next intel
> > driver upload; I filed a bug to track this some days ago.)
>
> i think yer talking about #555906, right? AFAICT, that patch looks to
Yep.
> just by-default add the kms option to the 915 driver. But given that
> the unstable kernel itself doesn't support KMS, this doesn't seem like
> it would make a difference:
>
> 0 dkg@pip:~$ grep KMS /boot/config-2.6.31-1-686
> # CONFIG_DRM_I915_KMS is not set
> # CONFIG_DRM_RADEON_KMS is not set
> 0 dkg@pip:~$
>
> or would it? I haven't fully gotten my head around all the
> relationships and implications of KMS (yet another reason i don't think
> i'm really qualified to be on the debian X team).
>
These options control the default value for the modeset parameter. Even
when not set the drivers are kms-capable (at least i915, and that will
be true for radeon as well in 2.6.32 IIRC).
> Anyway, i have intel hardware that i use regularly, and i'd be willing
> to test out builds as a guinea pig. I can't really take on X in general
> though, i just don't know enough about the grubby details.
>
The problem right now is not so much the lack of testers as the lack of
people willing (and having time) to take care of the packages, meaning
mostly packaging new upstream releases and dealing with bugs.
But thanks for your interest :)
Cheers,
Julien
Reply to: