Re: [PATCH] dt-bindings: display: bridge: lt9211: Require data-lanes on DSI input ports

[Marek Vasut <marex@nabladev.com>]

On 4/7/26 10:00 AM, Krzysztof Kozlowski wrote:

>> NOTE: For example Linux kernel driver does already use that information
>>        and fails to probe if it is missing. There are currently no intree
> 
> The first sentence must be part of the commit msg. That is important
> reason why you are doing this... but I don't see how you achieve any of
> this. Look:
> 
> 
>>        users for this binding, so no new warnings will be generated once
>>        this is applied, but a new user is about to be added.
> 
> What warnings? How?

There are no in-tree users of this binding, so no DT checker warnings 
will be produced on existing in-tree DTs. I am in the process of adding 
a DTO which uses this binding now in arm64: dts: imx8mm: imx8mp: Add 
DTOs for Data Modul i.MX8M Mini and Plus eDM SBC

>> ---
>>   .../display/bridge/lontium,lt9211.yaml        | 37 ++++++++++++++++++-
>>   1 file changed, 35 insertions(+), 2 deletions(-)
>>
>> diff --git a/Documentation/devicetree/bindings/display/bridge/lontium,lt9211.yaml b/Documentation/devicetree/bindings/display/bridge/lontium,lt9211.yaml
>> index 9a6e9b25d14a9..5264fb2b68b78 100644
>> --- a/Documentation/devicetree/bindings/display/bridge/lontium,lt9211.yaml
>> +++ b/Documentation/devicetree/bindings/display/bridge/lontium,lt9211.yaml
>> @@ -36,18 +36,50 @@ properties:
>>   
>>       properties:
>>         port@0:
>> -        $ref: /schemas/graph.yaml#/properties/port
>> +        $ref: /schemas/graph.yaml#/$defs/port-base
> 
> OK, that's correct.
> 
>> +        unevaluatedProperties: false
>>           description:
>>             Primary MIPI DSI port-1 for MIPI input or
>>             LVDS port-1 for LVDS input or DPI input.
>>   
>> +        properties:
>> +          endpoint:
>> +            $ref: /schemas/media/video-interfaces.yaml#
>> +            unevaluatedProperties: false
> 
> That's correct.
> 
>> +
>> +            properties:
>> +              data-lanes:
>> +                description: array of physical DSI data lane indexes.
>> +                minItems: 1
>> +                items:
>> +                  - const: 1
>> +                  - const: 2
>> +                  - const: 3
>> +                  - const: 4
> 
> That's almost redundant in this context - it was already there - and the
> point is that it solves noting in the problem you had. Binding still
> does not validate the ABI and does not match it, still.
> 
> Since commit foo bar, driver needs data-lanes, so what you need to do is
> allow them and to require them. You can also specify their constraints
> if device can be configured multiple ways, up to 4 lanes.
Please pardon my ignorance, what exactly do you propose I change in this 
patch ?