Re: [PATCH 1/3] dt-bindings: display: panel: Rename WL-355608-A8 panel

[Neil Armstrong <neil.armstrong@linaro.org>]

On 30/08/2024 03:43, Ryan Walklin wrote:
> 
> 
>> On 28 Aug 2024, at 7:07 PM, Maxime Ripard <mripard@kernel.org> wrote:
>>
>> On Tue, Aug 27, 2024 at 06:28:21PM GMT, Neil Armstrong wrote:
>>>> Thanks both for the further feedback, agreed logical to use the
>>>> device vendor and panel serial number, ie "anbernic,wl-355608-a8".
>>>> Will post a V2 with a comment to that effect.
>>>
>>> Well in this case we can keep "wl-355608-a8", because the panel vendor
>>> _is not_ anbernic.
>>
>> And it's not a generic or ubiquitous device either. We've been over
>> this already, anbernic is the best we have.
> 
> I don’t have a strong preference either way but agree the anbernic vendor string is the best compromise.

The only valid compatible with anbernic would be to use the exact device in use and not a wildcard,
so you said the 3 devices using this panel are:
  anbernic,rg35xx-2024
  anbernic,rg35xx-plus
  anbernic,rg35xx-h
you should introduce 3 compatibles:
  anbernic,rg35xx-2024-panel
  anbernic,rg35xx-plus-panel
  anbernic,rg35xx-h-panel

but it's duplicating for nothing, to you should use fallbacks for 2 of them to have in DT :

  anbernic,rg35xx-2024-panel
  anbernic,rg35xx-plus-panel, anbernic,rg35xx-2024-panel
  anbernic,rg35xx-h-panel, anbernic,rg35xx-2024-panel

and only use anbernic,rg35xx-2024-panel in the driver.

In this case bindings should be like:
   properties:
     compatible:
       oneOf:
         - const: anbernic,rg35xx-2024-panel
         - items:
             - enum:
                 - anbernic,rg35xx-plus-panel
                 - anbernic,rg35xx-h-panel
             - const: anbernic,rg35xx-2024-panel

(of course I selected rg35xx-2024 as the primary one, it could be another, usually the older one)

Neil

> 
> Regards,
> 
> Ryan