On Sun, Nov 28, 2010 at 08:05:06PM +0100, Guennadi Liakhovetski wrote:
quoted text
> On Sun, 28 Nov 2010, Alberto Panizzo wrote:

quoted text
> > In certain machines, camera devices are supplied directly
> > by a number of regulators. This patch add the ability to drive
> > these regulators directly by the soc_camera driver.

quoted text
> IIRC, there has been a discussion a while ago, how to supply power to 
> cameras by regulators. Someone has tried to provide a .power() hook in the 
> platform code, but the problem was the order of driver loading / probing. 
> So, how about doing the following:

quoted text
> 1. in your platform code you register a notifier like
> 	bus_register_notifier(&soc_camera_bus_type, &cam_notifier);

FWIW I'm looking at implementing a standard regulator API feature along
these lines in the background.  This should hopefully mean we don't need
driver support for most simple power control applications.  No ETA yet.

quoted text
> The reasons why I do not want to add this to the core are: (1) I do not 
> want to have two methods for turning power on and off: a platform provided 
> .power() hook and and a set of regulators, (2) would anyone really want to 
> use several regulators for a camera sensor? If so, can it be the case, 
> that, for example, the regulators have to be switched off in the reverse 
> order to switching on? Or something else? (3) regulators can often do 
> more, than just set one of two power levels - for on and off. What if a 
> need arises to use other voltages?

The way MMC handled this was to provide a standard version of the hook
in the core which could be used by platforms with regulators supplying
the device - they just assign the appropriate function as their power()
operation AIUI.  That seems a fairly clean way of keeping stuff in the
core without giving up the flexibility.

quoted text
> Is there any really good reason, why we _have_ to do this in soc-camera 
> core?

It does save everyone open coding stuff.
--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/