On Mon, Jan 13, 2014 at 9:37 PM, Josh Boyer <jwboyer(a)fedoraproject.org> wrote:
On Mon, Jan 13, 2014 at 4:23 PM, Nicolas Chauvet
<kwizart(a)gmail.com> wrote:
> Hello,
>
> I need few commits scheduled for 3.14 to be backported into 3.13 Fedora
> kernel.
As helpful as we try to be, we are not a backporting service. If you
would like to see something backported to a particular Fedora release
or kernel version, the best way to do that would be to work with the
Fedora ARM group and send patches. Normally that means you're also
signing up to provide the support for the patches you're submitting.
> This is needed to fully support the Toshiba AC100 device on Fedora ARM.
> I haven't found the ref for some swarren commit, but they are marked as
> APPLIED, so they will probably hit linux-next soon.
>
> - Select the right RTC
> linux-next: 553c0a200e2082686fd9b829b77f7df6ebae14e9
>
> - Simple-panel driver
> linux-next: 280921de7241ee63184c8baa89ec3fe0122dedb3
> linux-next: 210fcd9d9cf1ad6ebfae3b46b457e602c8f8cdc2
>
http://nv-tegra.nvidia.com/gitweb/?p=user/swarren/linux-2.6.git;a=commitd...
>
http://nv-tegra.nvidia.com/gitweb/?p=user/swarren/linux-2.6.git;a=commitd...
>
> Adds: DRM_PANEL boot and DRM_PANEL_SIMPLE tristate
>
I'm hesitant to add an entirely new driver as a backport. Not hugely
so, but given that it's only been in linux-next I have no idea how
much it's been tested.
> - Misc fix
>
http://nv-tegra.nvidia.com/gitweb/?p=user/swarren/linux-2.6.git;a=commitd...
>
> I plan to use a rawhidenodebug repo to compose a fedora 20 remix.
> Then hopefully everything should be in for fedora 21.
Is there a reason you can't wait for F21 and a 3.14 kernel then?
We're doing something similar to what you're requesting with the
AllWinner devices that Hans has been supporting but waiting for the
3.14 merge window to start. I'd strongly suggest going that route and
working with the ARM SIG as they're in the middle of working out their
hardware support story.
I would prefer we wait until it lands upstream. If it was a device
which was widely available such as the BeagleBone or Wandboard I would
consider it but unfortunately the AC100 is a somewhat corner case
device which I personally don't believe is worth the pain and can be
maintained via a remix. Don't get me wrong, I would love to support
this device upstream eventually, I have one myself, but if we start
doing it for a device here and a device there we're suddenly carrying
100 patches before we know it and it's unmaintainable.
Peter