Re: DRM based modesetting for drm-next

2008-10-30 Thread Jesse Barnes
On Thursday, October 30, 2008 2:03 pm Jesse Barnes wrote:
> In the hopes of getting the kernel mode setting stuff in shape for the
> 2.6.29 merge window I've been trying to get things into shape.  This patch
> set is a very rough start at getting things merge-ready.  There's still a
> lot left to do:
>   - document the new interfaces (basically finish the DRM developer guide I
> started a long time ago), should help get other drivers ported
>   - fix i915 interrupt support, this patch set basically reverts most of
> the earlier mode setting changes; there are lots of fixes upstream I didn't
> want to lose
>   - fix bugs, there are lots of these to go around (right now load/unload
> and lastclose seem to cause massive memory corruption)
>   - test the new code with modeset=0 against old drivers
>
> These patches sit on top of my GTT mapping patches (posted earlier) since a
> KMS aware 2D driver really wants GTT mapping too.  Also I just pushed a fix
> to the 2D driver that, in combination with my EXA pixmap management patches
> (which in turn depend on the libdrm bits), gets it limping along with
> kernel mode setting support.
>
> My rough list of priorities is as follows:
>   1) make load/unload solid (this will make subsequent development *much*
>  easier)
>   2) make X startup/teardown solid (this is probably a dup of (1) at this
>  point)
>   3) fixup the IRQ handling stuff, re-adding the ripped out hot plug stuff
>   4) minimize the diff with upstream (reverting some of the cleanups I
>  foolishly included in the initial mode setting work)
>   5) change the panic/oops stuff to use oops_begin rather than the panic
>  notifier
> ...and of course fix bugs as they're discovered along the way.  I should
> have some patches early next week for our QA team and the broader community
> to bang on, but any feedback people have for this set can be included in
> the next post, so feel free to send me feedback.

Wow they turned out to be fairly huge.  Definitely need to be split up some 
more.

In the likely event that the mailing list fails to deliver the 3 patches in 
reply to this message, you can grab them from kernel.org:  
http://www.kernel.org/pub/linux/kernel/people/jbarnes/patches/drm-kms.

Thanks,
Jesse

-
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
--
___
Dri-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dri-devel


DRM based modesetting for drm-next

2008-10-30 Thread Jesse Barnes
In the hopes of getting the kernel mode setting stuff in shape for the 2.6.29 
merge window I've been trying to get things into shape.  This patch set is a 
very rough start at getting things merge-ready.  There's still a lot left to 
do:
  - document the new interfaces (basically finish the DRM developer guide I
started a long time ago), should help get other drivers ported
  - fix i915 interrupt support, this patch set basically reverts most of the
earlier mode setting changes; there are lots of fixes upstream I didn't
want to lose
  - fix bugs, there are lots of these to go around (right now load/unload and
lastclose seem to cause massive memory corruption)
  - test the new code with modeset=0 against old drivers

These patches sit on top of my GTT mapping patches (posted earlier) since a 
KMS aware 2D driver really wants GTT mapping too.  Also I just pushed a fix to 
the 2D driver that, in combination with my EXA pixmap management patches 
(which in turn depend on the libdrm bits), gets it limping along with kernel 
mode setting support.

My rough list of priorities is as follows:
  1) make load/unload solid (this will make subsequent development *much*
 easier)
  2) make X startup/teardown solid (this is probably a dup of (1) at this
 point)
  3) fixup the IRQ handling stuff, re-adding the ripped out hot plug stuff
  4) minimize the diff with upstream (reverting some of the cleanups I
 foolishly included in the initial mode setting work)
  5) change the panic/oops stuff to use oops_begin rather than the panic
 notifier
...and of course fix bugs as they're discovered along the way.  I should have 
some patches early next week for our QA team and the broader community to bang 
on, but any feedback people have for this set can be included in the next 
post, so feel free to send me feedback.

Thanks,
-- 
Jesse Barnes, Intel Open Source Technology Center


-
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
--
___
Dri-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dri-devel