Re: [PATCH v3 8/8] dt-bindings: Add rt5033 mfd, regulator and charger

[Krzysztof Kozlowski]

On 01/05/2023 23:16, Jakob Hauser wrote:
> Hi Krzysztof,
> 
> On 01.05.23 19:35, Jakob Hauser wrote:
>> On 01.05.23 09:21, Krzysztof Kozlowski wrote:
>>> On 28/04/2023 01:30, Jakob Hauser wrote:
>>>> Add device tree binding documentation for rt5033 multifunction 
>>>> device, voltage
>>>> regulator and battery charger.
>>>>
>>>> Cc: Beomho Seo <beomho.seo@xxxxxxxxxxx>
>>>> Cc: Chanwoo Choi <cw00.choi@xxxxxxxxxxx>
>>>> Signed-off-by: Jakob Hauser <jahau@xxxxxxxxxxxxxx>
>>>
>>>
>>> (...)
>>>
>>>> +
>>>> +required:
>>>> +  - monitored-battery
>>>> +
>>>> +additionalProperties: false
>>>> +
>>>> +examples:
>>>> +  - |
>>>> +    charger {
>>>> +        compatible = "richtek,rt5033-charger";
>>>> +        monitored-battery = <&battery>;
>>>> +        extcon = <&muic>;
>>>
>>>
>>> Everything up to here looked ok, but extcon is not a hardware property.
>>> Please do not mix adding missing bindings for existing device with
>>> adding new properties. You should use connector for the USB port.
>>>
> 
> ...
> 
>> And how to set up the rt5033-charger to retrieve the information of the 
>> extcon/muic driver in that case?
>>
> 
> ...
> 
> To add more context:
> According to my understanding, the extcon subsystem provides three ways 
> to get an extcon device [3]:
> - by name
> - by devicetree node
> - by phandle
> 
> For rt5033-charger, the extcon device name can be different depending on 
> the consumer device. For the node I wouldn't know how to get from the 
> charger/mfd node to the extcon node, I don't see a direct relation in 
> case of rt5033-charger (it's no parent node or something like that). 
> Therefore I chose the third option: phandle.
> 
> In the rt5033-charger driver, the location of the 
> extcon_get_edev_by_phandle() call is shown in link [4], it gets added in 
> patch 6.
> 

Hi Jakob,

I am currently busy, so I won't be able to help you and dig in your
reply. I will get to you a bit later (or maybe Rob will help here).

However please check if ports graph does not solve your case:
https://lore.kernel.org/all/20230501121111.1058190-6-bryan.odonoghue@xxxxxxxxxx/

It is already used for orientation and usb-role-switch (which should
solve the need for extcon, AFAIR).

If it does not help, ping me again, and I'll try to get to you a bit later.

Apologies for this, just very busy times. :)

Best regards,
Krzysztof