Re: [PATCH 1/4] dt-bindings: brige: lt9611c: add port-select property for LT9611C

[Krzysztof Kozlowski <krzk@kernel.org>]

On 22/06/2026 14:08, Laurent Pinchart wrote:
> On Mon, Jun 22, 2026 at 05:31:36PM +0530, Mohit Dsor wrote:
>> On Thu, Jun 11, 2026 at 12:40:38PM +0200, Krzysztof Kozlowski wrote:
>>> On Thu, Jun 11, 2026 at 02:44:56AM +0530, Mohit Dsor wrote:
>>>> Add a new optional `lontium,port-select` property to describe the DSI
>>>> input port configuration for the LT9611C variant, which supports
>>>> single-port (A or B) and dual-port (A+B) operation.
>>>>
>>>> This property allows explicitly selecting the active DSI input port(s):
>>>>   0 = port A (default)
>>>>   1 = port B
>>>>   2 = ports A and B (dual-port)
>>>>
>>>> Signed-off-by: Mohit Dsor <mohit.dsor@oss.qualcomm.com>
>>>> ---
>>>>  .../devicetree/bindings/display/bridge/lontium,lt9611.yaml  | 13 +++++++++++++
>>>>  1 file changed, 13 insertions(+)
>>>>
>>>> diff --git a/Documentation/devicetree/bindings/display/bridge/lontium,lt9611.yaml b/Documentation/devicetree/bindings/display/bridge/lontium,lt9611.yaml
>>>> index e0821a63d9d7..77220f893bf8 100644
>>>> --- a/Documentation/devicetree/bindings/display/bridge/lontium,lt9611.yaml
>>>> +++ b/Documentation/devicetree/bindings/display/bridge/lontium,lt9611.yaml
>>>> @@ -41,6 +41,17 @@ properties:
>>>>    vcc-supply:
>>>>      description: Regulator for 3.3V IO power.
>>>>  
>>>> +  lontium,port-select:
>>>> +    $ref: /schemas/types.yaml#/definitions/uint32
>>>> +    enum: [0, 1, 2]
>>>> +    default: 0
>>>> +    description: |
>>>> +      Selects which DSI input port(s) the bridge uses. Only relevant for
>>>> +      the lontium,lt9611c compatible.
>>>> +        0 = PORT_SELECT_A  - single DSI port A (default)
>>>> +        1 = PORT_SELECT_B  - single DSI port B
>>>> +        2 = PORT_SELECT_AB - dual DSI ports A and B
>>>
>>> Why graph is not enough? Seems exactly duplicating the graph ports.
>>>
>>> Best regards,
>>> Krzysztof
>>>
>> Hi Krzysztof,
>>
>> Thanks for the review.
>>
>> The graph describes the physical connectivity between endpoints,
>> however it does not fully capture the internal mode of operation of
>> the LT9611C. This variant supports multiple functional configurations
>> (single-port A, single-port B, or dual-port A+B), which affect how the
>> hardware internally combines or selects DSI inputs.
>>
>> In particular:
>> - The graph can describe connections to both ports, but it does not
>>   indicate whether the device should operate in single-port or dual-port
>>   aggregation mode.
>> - For single-port use, both ports may be described in DT for board
>>   consistency, while the driver still needs to know which port is
>>   actively selected.
>> - Dual-port mode requires explicit configuration even when both
>>   endpoints are present in the graph.
> 
> If both modes of operation are possible on a given board, then it sounds
> like the mode should be selected at runtime, not hardcoded in the device
> tree.

Yeah. Especially "while the driver still needs to know which port is
actively selected" for a case when DT already clearly defines which
ports are available, feels like 100% runtime decision, e.g. based on
what is actually plugged into the connector.

Best regards,
Krzysztof