linux-backports-input proposal

[Stefan Bader]

On 07/14/2010 03:18 PM, Stefan Bader wrote:
> On 07/14/2010 03:12 PM, Chase Douglas wrote:
>> On Wed, 2010-07-14 at 04:57 -0500, Taylor LeMasurier-Wren wrote:
>>> Dear Kernel Team,
>>> In Lucid and Maverick, the Wacom Bamboo Pen & Touch series, a newer
>>> series of wacom tablets, is broken and doesn't work. The current
>>> working solution is to download the upstream linuxwacom source and
>>> rebuild the wacom kernel module. As far as I can tell, the userspace
>>> driver is fine, and doesn't need to be updated. Right now, I've been
>>> recommending people who have this problem to download my wacom-dkms
>>> package from Wacom Kernel Module PPA (ppa:ripps818/wacom). This
>>> solution is easier than having them compile their own module, but it
>>> would be better to have a more official solution. Smb on
>>> #ubuntu-kernel recommended a l-b-m package, and suggested I contact
>>> this mailing list for more information. A wacom module backport
>>> package seems to be a more ideal solution to the problem. I'd like to
>>> know if it would be possible to implement one in Maverick, and a
>>> backport of it to Lucid?
>>
>> First, is the issue fixed in upstream linux yet? If not, it needs to be
>> fixed there first. The wacom stuff is kinda weird cause they are taking
>> the opposite approach from everyone else and moving their drivers out of
>> the kernel instead of the other way around.
>>
>> If a fix has been merged by upstream, we can take patches in Maverick to
>> have the issue fixed in Ubuntu as well.
>>
>> As for adding a new l-b-m input package, it may be worthwhile. However,
>> we can't add packages retroactively to a released distro (at least not
>> yet). Thus, we can't create a new l-b-m-input package for Lucid. I would
>> be all for creating a new package for Maverick though.
>>
> 
> This is not completely true. It is possible to do a MIR (though listening to Tim
> sounds like this is a painful thing). What I am not completely clear is whether
> this is required for the case of just having a source package which is already
> there just produce a new binary sub-package (mainly to have things a bit better
> sorted). So l-b-m would just produce a l-b-m-input additionally. Like it
> currently produces l-b-m-wireless and l-b-m-alsa....
> 
> Overall for Maverick and the future it would be the best if that driver update
> would somehow go to upstream. But its too late for 2.6.35...
> So we might need backports for Lucid and Maverick.
> 
> -Stefan
>> -- Chase
>>
>>
>>
> 
> 
So I verified that there is no restriction on adding new binary packages that
are produced by the l-b-m package. So the next steps would be to have opened a
bug report against the linux-backports-modules-2.6.35 package that explains why
this is required (not in upstream linux but required for ...) and point to the
sources required to backport.
Taylor, would you be able to go ahead with that. Then point us to the bug created?

-Stefan