Re: [PATCH 1/2] include: dt-binding: input: create a DT header defining key codes.

[Stephen Warren <swarren@wwwdotorg.org>]

On 08/12/2013 11:15 AM, Dmitry Torokhov wrote:
> On Thu, Aug 08, 2013 at 11:25:27AM -0600, Stephen Warren wrote:
>> On 08/08/2013 11:13 AM, Dmitry Torokhov wrote:
>>> On Thu, Aug 08, 2013 at 10:41:20PM +0530, Laxman Dewangan wrote:
>>>> On Wednesday 07 August 2013 12:58 AM, Stephen Warren wrote:
>>>>> (CC'ing DT bindings maintainers too, hence quoting a bit of the patch)
>>>>>
>>>>> On 08/06/2013 08:12 AM, Laxman Dewangan wrote:
>>>>>> Many of Key device tree bindings uses the constant number as key code
>>>>>> which matches with kernel header key code and then comment as follows
>>>>>> for reference/better readability:
>>>>>> 	linux,code = <102>; /* KEY_HOME */
>>>>>>
>>>>>> Create a DT header which defines all the key code so that DT key bindings
>>>>>> can use it as follows:
>>>>>> 	linux,code = <KEY_HOME>;
>> ...
>>> I am uncomfortable that we are duplicating the whole list of keycodes
>>> here. Do you think we could split keycodes and other event codes from
>>> include/uapi/linux/input.h into include/uapi/linux/input-events.h and
>>> unclude it directly into DTS?
>>
>> Device tree files can only include headers in include/dt-bindings.
>> Eventually/soon, all the DT files and the headers in the dt-bindings/
>> directory will be split out into a separate git tree, which does not
>> have access to include/uapi, or indeed any other Linux kernel header files.
> 
> OK, that makes it easier for me - please merge through whatever tree
> that owns include/dt-bindings.

Are there any objections to my taking both patches through the Tegra
tree? Patch 2/2 depends on 1/2 and is Tegra-specific. This should work
out fine unless anyone else wants to use the new header this kernel
cycle; if they do, this should go into a topic branch in the DT tree
that I can pull into the Tegra tree.