[RFC] adp1653: Add device tree bindings for LED controller

[pali.rohar@gmail.com (Pali Rohár)]

On Monday 17 November 2014 16:04:07 Sakari Ailus wrote:
> Hi Pali,
> 
> On Mon, Nov 17, 2014 at 04:01:31PM +0100, Pali Roh?r wrote:
> > On Monday 17 November 2014 15:55:46 Tony Lindgren wrote:
> > > * Pavel Machek <pavel@ucw.cz> [141117 02:17]:
> > > > On Mon 2014-11-17 11:09:45, Pali Roh?r wrote:
> > > > > On Monday 17 November 2014 11:05:19 Pavel Machek wrote:
> > > > > > Hi!
> > > > > > 
> > > > > > On Mon 2014-11-17 09:43:19, Pali Roh?r wrote:
> > > > > > > On Sunday 16 November 2014 08:59:28 Pavel Machek 
wrote:
> > > > > > > > For device tree people: Yes, I know I'll have to
> > > > > > > > create file in documentation, but does the
> > > > > > > > binding below look acceptable?
> > > > > > > > 
> > > > > > > > I'll clean up driver code a bit more, remove the
> > > > > > > > printks. Anything else obviously wrong?
> > > > > > > 
> > > > > > > I think that this patch is probably not good and
> > > > > > > specially not for n900. adp1653 should be
> > > > > > > registered throw omap3 isp camera subsystem which
> > > > > > > does not have DT support yet.
> > > > > > 
> > > > > > Can you explain?
> > > > > > 
> > > > > > adp1653 is independend device on i2c bus, and we
> > > > > > have kernel driver for it (unlike rest of n900
> > > > > > camera system). Just now it is unusable due to lack
> > > > > > of DT binding. It has two functions, LED light and
> > > > > > a camera flash; yes, the second one should be
> > > > > > integrated to the rest of camera system, but that
> > > > > > is not yet merged. That should not prevent us from
> > > > > > merging DT support for the flash, so that this part
> > > > > > can be tested/maintained.
> > > > > 
> > > > > Ok. When ISP camera subsystem has DT support somebody
> > > > > will modify n900 DT to add camera flash from adp1653
> > > > > to ISP... I believe it will not be hard.
> > > > 
> > > > Exactly. And yes, I'd like to get complete camera
> > > > support for n900 merged. But first step is "make sure
> > > > existing support does not break".
> > > 
> > > There's nothing stopping us from initializing the camera
> > > code from pdata-quirks.c for now to keep it working.
> > > Certainly the binding should be added to the driver, but
> > > that removes a dependency to the legacy booting mode if
> > > things are otherwise working.
> > > 
> > > Regards,
> > > 
> > > Tony
> > 
> > Tony, legacy board code for n900 is not in mainline tree.
> > And that omap3 camera subsystem for n900 is broken since
> > 3.5 kernel... (both Front and Back camera on n900 show only
> > green picture).
> 
> Can you capture raw bayer images correctly? I assume green
> means YUV buffers that are all zero.
> 
> Do you know more specifically which patch breaks it?

CCing freemangordon (Ivaylo Dimitrov). He tried to debug it 
months ago but without success. Should know more info about this 
problem.

I think that commit which broke it was not bisected...

-- 
Pali Roh?r
pali.rohar at gmail.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20141117/878e3073/attachment.sig>