Date:	Fri, 16 Jan 2015 10:07:28 +0100
From:	Jacek Anaszewski <j.anaszewski@...sung.com>
To:	Rob Herring <robherring2@...il.com>
Cc:	linux-leds@...r.kernel.org,
	"linux-media@...r.kernel.org" <linux-media@...r.kernel.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"devicetree@...r.kernel.org" <devicetree@...r.kernel.org>,
	Kyungmin Park <kyungmin.park@...sung.com>,
	Bartlomiej Zolnierkiewicz <b.zolnierkie@...sung.com>,
	Pavel Machek <pavel@....cz>, Bryan Wu <cooloney@...il.com>,
	Richard Purdie <rpurdie@...ys.net>, sakari.ailus@....fi,
	Sylwester Nawrocki <s.nawrocki@...sung.com>,
	Rob Herring <robh+dt@...nel.org>,
	Pawel Moll <pawel.moll@....com>,
	Mark Rutland <mark.rutland@....com>,
	Ian Campbell <ijc+devicetree@...lion.org.uk>,
	Kumar Gala <galak@...eaurora.org>,
	Liam Girdwood <lgirdwood@...il.com>,
	Mark Brown <broonie@...nel.org>
Subject: Re: [PATCH/RFC v10 03/19] DT: leds: Add led-sources property

On 01/15/2015 03:24 PM, Rob Herring wrote:
> On Tue, Jan 13, 2015 at 2:42 AM, Jacek Anaszewski
> <j.anaszewski@...sung.com> wrote:
>> On 01/12/2015 05:55 PM, Rob Herring wrote:
>>>
>>> Adding Mark B and Liam...
>>>
>>> On Mon, Jan 12, 2015 at 10:10 AM, Jacek Anaszewski
>>> <j.anaszewski@...sung.com> wrote:
>>>>
>>>> On 01/12/2015 02:52 PM, Rob Herring wrote:
>>>>>
>>>>>
>>>>> On Mon, Jan 12, 2015 at 2:32 AM, Jacek Anaszewski
>>>>> <j.anaszewski@...sung.com> wrote:
>>>>>>
>>>>>> On 01/09/2015 07:33 PM, Rob Herring wrote:
>>>>>>>
>>>>>>> On Fri, Jan 9, 2015 at 9:22 AM, Jacek Anaszewski
>>>>>>> <j.anaszewski@...sung.com> wrote:
>>>>>>>>
>>>>>>>> Add a property for defining the device outputs the LED
>>>>>>>> represented by the DT child node is connected to.
>>>
>>>
>>> [...]
>>>
>>>>>>>> b/Documentation/devicetree/bindings/leds/common.txt
>>>>>>>> index a2c3f7a..29295bf 100644
>>>>>>>> --- a/Documentation/devicetree/bindings/leds/common.txt
>>>>>>>> +++ b/Documentation/devicetree/bindings/leds/common.txt
>>>>>>>> @@ -1,6 +1,10 @@
>>>>>>>>      Common leds properties.
>>>>>>>>
>>>>>>>>      Optional properties for child nodes:
>>>>>>>> +- led-sources : Array of bits signifying the LED current regulator
>>>>>>>> outputs the
>>>>>>>> +               LED represented by the child node is connected to (1
>>>>>>>> -
>>>>>>>> the LED
>>>>>>>> +               is connected to the output, 0 - the LED isn't
>>>>>>>> connected
>>>>>>>> to the
>>>>>>>> +               output).
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> Sorry, I just don't understand this.
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> In some Flash LED devices one LED can be connected to one or more
>>>>>> electric current outputs, which allows for multiplying the maximum
>>>>>> current allowed for the LED. Each sub-LED is represented by a child
>>>>>> node in the DT binding of the Flash LED device and it needs to declare
>>>>>> which outputs it is connected to. In the example below the led-sources
>>>>>> property is a two element array, which means that the flash LED device
>>>>>> has two current outputs, and the bits signify if the LED is connected
>>>>>> to the output.
>>>>>
>>>>>
>>>>>
>>>>> Sounds like a regulator for which we already have bindings for and we
>>>>> have a driver for regulator based LEDs (but no binding for it).
>>>>
>>>>
>>>>
>>>> Do you think of drivers/leds/leds-regulator.c driver? This driver just
>>>> allows for registering an arbitrary regulator device as a LED subsystem
>>>> device.
>>>>
>>>> There are however devices that don't fall into this category, i.e. they
>>>> have many outputs, that can be connected to a single LED or to many LEDs
>>>> and the driver has to know what is the actual arrangement.
>>>
>>>
>>> We may need to extend the regulator binding slightly and allow for
>>> multiple phandles on a supply property, but wouldn't something like
>>> this work:
>>>
>>> led-supply = <&led-reg0>, <&led-reg1>, <&led-reg2>, <&led-reg3>;
>>>
>>> The shared source is already supported by the regulator binding.
>>
>>
>> I think that we shouldn't split the LED devices into power supply
>> providers and consumers as in case of generic regulators. From this
>> point of view a LED device current output is a provider and a discrete
>> LED element is a consumer. In this approach each discrete LED element
>> should have a related driver which is not how LED devices are being
>> handled in the LED subsystem, where there is a single binding for a LED
>> device and there is a single driver for it which creates separate LED
>> class devices for each LED connected to the LED device output. Each
>> discrete LED is represented by a child node in the LED device binding.
>>
>> I am aware that it may be tempting to treat LED devices as common
>> regulators, but they have their specific features which gave a
>> reason for introducing LED class for them. Besides, there is already
>> drivers/leds/leds-regulator.c driver for LED devices which support only
>> turning on/off and setting brightness level.
>>
>> In your proposition a separate regulator provider binding would have
>> to be created for each current output and a separate binding for
>> each discrete LED connected to the LED device. It would create
>> unnecessary noise in a dts file.
>>
>> Moreover, using regulator binding implies that we want to treat it
>> as a sheer power supply for our device (which would be a discrete LED
>> element in this case), whereas LED devices provide more features like
>> blinking pattern and for flash LED devices - flash timeout, external
>> strobe and flash faults.
>
> Okay, fair enough. Please include some of this explanation in the
> binding description.
>
> I do still have some concerns about led-sources and whether it can
> support other scenarios. It is very much tied to the parent node. Are
> there any cases where we don't want the LEDs to be sub nodes? Perhaps
> the LEDs are on a separate daughterboard from the driver/supply and we
> can have different drivers. It's a stretch maybe.

I think it is. Such arrangements would introduce problems also to the
other existing bindings. Probably not only LED subsystem related ones.

> Or are there cases
> where you need more information than just the connection?

Currently I can't think of any.

Modified rough proposal of the description:


-Optional properties for child nodes:
+LED and flash LED devices provide the same basic functionality as
+current regulators, but extended with LED and flash LED specific 
+features like blinking patterns, flash timeout, flash faults and
+external flash strobe mode.
+
+Many LED devices expose more than one current output that can be
+connected to one or more discrete LED component. Since the arrangement
+of connections can influence the way of the LED device initialization,
+the LED components have to be tightly coupled with the LED device
+binding. They are represented in the form of its child nodes.
+
+Optional properties for child nodes (if a LED device exposes only one
+current output the properties can be placed directly in the LED device
+node):
+- led-sources : Array of connection states between all LED current
+               sources exposed by the device and this LED (1 - this LED
+               is connected to the current output with index N, 0 -
+               this LED isn't connected to the current output with
+               index N); the mapping of N-th element of the array to
+               the physical device output should be defined in the LED
+               driver binding.


-- 
Best Regards,
Jacek Anaszewski
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Powered by blists - more mailing lists