On 2.05.2023 13:16, Bryan O'Donoghue wrote:
> On 02/05/2023 12:13, Konrad Dybcio wrote:
>>
>>
>> On 2.05.2023 13:03, Bryan O'Donoghue wrote:
>>> On 02/05/2023 12:00, Konrad Dybcio wrote:
>>>>> + #address-cells = <1>;
>>>>> + #size-cells = <0>;
>>>>> +
>>>>> + port@0 {
>>>>> + reg = <0>;
>>>>> + pm8150b_role_switch_out: endpoint {
>>>> Similarly to the QMPPHY, the port definition can be moved to
>>>> the common node in the SoC DTSI
>>>
>>> But then the connector would have to be defined in the SoC dtsi and not all derivatives of SoC can be assumed to have a usb-c-connector.
>> Not quite!
>>
>> You can define an empty endpoint (like we do with DSI<->panel ones) and
>> fill it in on the device side.
>
> Sorry you're saying to define as an example the port here in the dtsi
>
> &usb_1_dwc3 {
> dr_mode = "otg";
> usb-role-switch;
===
> port {
> dwc3_role_switch_in: endpoint {
> remote-endpoint = <&pm8150b_role_switch_out>;
> };
> };
===
this part (minus remote-endpoint) would go to SoC dtsi
remote-endpoint would be assigned on a per-device basis
> };
>
> and to leave the reciprocal definition in the connector to the dts ?
>
> &pm8150b_typec {
>
====
> connector {
> compatible = "usb-c-connector";
> ports {
> #address-cells = <1>;
> #size-cells = <0>;
>
> port@0 {
> reg = <0>;
> pm8150b_role_switch_out: endpoint {
> remote-endpoint = <&dwc3_role_switch_in>;
> };
> };
====
this part (minus remote-endpoint) would go to pm8150b.dtsi
remote-endpoint would be assigned (or left empty) on a per-device basis
Konrad
>
