Re: [PATCH 2/3] devicetree: bindings: use input-event-codes.h for evdev codes

[Hans de Goede <hdegoede@redhat.com>]

Hi,

On 09/16/2015 10:40 AM, Ian Campbell wrote:
> On Wed, 2015-09-16 at 10:27 -0400, Hans de Goede wrote:
>> Hi,
>>
>> On 09/15/2015 07:24 AM, Ian Campbell wrote:
>>> On Sat, 2015-09-12 at 15:26 +0200, Hans de Goede wrote:
>>>> Add a symlink to uapi/linux/input-event-codes.h, and include that
>>>> instead
>>>> of (re)defining all the evdev type and code values in
>>>> dt-bindings/input/input.h. This way we do not need to keep all the
>>>> event codes synced manually.
>>>>
>>>> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
>>>> ---
>>>>    include/dt-bindings/input/input-event-codes.h |   1 +
>>>>    include/dt-bindings/input/input.h             | 510 +--------------
>>>> -----
>>>> ------
>>>>    2 files changed, 2 insertions(+), 509 deletions(-)
>>>>    create mode 120000 include/dt-bindings/input/input-event-codes.h
>>>>
>>>> diff --git a/include/dt-bindings/input/input-event-codes.h
>>>> b/include/dt
>>>> -bindings/input/input-event-codes.h
>>>> new file mode 120000
>>>> index 0000000..693bbcd
>>>> --- /dev/null
>>>> +++ b/include/dt-bindings/input/input-event-codes.h
>>>> @@ -0,0 +1 @@
>>>> +../../uapi/linux/input-event-codes.h
>>>> \ No newline at end of file
>>>
>>> Hrm, I think this is going to cause trouble for the split DT repository
>>> at
>>> https://git.kernel.org/cgit/linux/kernel/git/devicetree/devicetree-reba
>>> sing.git/
>>>
>>> Since ../../ == include in both trees I think I can make it work by
>>> including include/uapi/linux in the conversion, either wholesale
>>> (easiest)
>>> or partially (potentially cleaner).
>>>
>>> Any opinions on that?
>>
>> The intend of the symlink was that the conversion script would copy the target,
>> rather then follow the symlink. This sorta assumes that there are will be not
>> symlinks under dt-bindings which link to files inside dt-bindings and thus
>> should be preserved as symlinks.
>>
>> If the copy done in the script will follow the symlink then nothing should
>> really change for the split DT repo.
>
> Interesting idea. I'll see if I can make the conversion routine do that.
>
> TBH I have a horrid feeling that this is going to be beyond git rewrite
> -branch, at least in the mode it is used in today.
>
>>> In principal I ought to regenerate the tree from scratch such that
>>> include/uapi/linux was "always" there and has history but I think we
>>> can
>>> live with it just appearing at some point in time wen the conversion
>>> script
>>> was updated?
>>>
>>> AIUI these values are required to only be used with properties whose
>>> names
>>> are of the form "linux,*", otherwise there would be strong complaints
>>> about
>>> include Linux uapi stuff into a generic device-tree binding.
>>
>> Correct, these values are only for linux,* properties, which is why I
>> initially put them in a file named dt-bindings/input/linux-event-codes.h
>
> I think that would be a better name for the symlink (but not the target of
> that symlink) and usage in the .dts*, given it's (hopefully) going to turn
> into a regular file in the exported repo.

Ok I can fix this in a next version of the patch-set (once we've figured
out how to deal with all this exactly).

Regards,

Hans