Re: [PATCH] dt-bindings: usb: Add binding for Microchip usb5744 hub controller

From: Krzysztof Kozlowski
Date: Thu May 04 2023 - 04:10:46 EST


On 04/05/2023 09:55, Michal Simek wrote:
>
>
> On 5/4/23 09:31, Krzysztof Kozlowski wrote:
>> CAUTION: This message has originated from an External Source. Please use proper judgment and caution when opening attachments, clicking links, or responding to this email.
>>
>>
>> On 04/05/2023 09:25, Michal Simek wrote:
>>>
>>>
>>> On 5/4/23 08:46, Krzysztof Kozlowski wrote:
>>>> On 03/05/2023 15:39, Michal Simek wrote:
>>>>> The Microchip usb5744 is a SS/HS USB 3.0 hub controller with 4 ports.
>>>>> The binding describes USB related aspects of the USB5744 hub, it as
>>>>> well cover the option of connecting the controller as an i2c slave.
>>>>> When i2c interface is connected hub needs to be initialized first.
>>>>> Hub itself has fixed i2c address 0x2D but hardcoding address is not good
>>>>> idea because address can be shifted by i2c address translator in the
>>>>> middle.
>>>>>
>>>>> Signed-off-by: Piyush Mehta <piyush.mehta@xxxxxxx>
>>>>> Signed-off-by: Michal Simek <michal.simek@xxxxxxx>
>>>>> ---
>>>>>
>>>>> It looks like that usb8041 has also an optional i2c interface which is not
>>>>> covered. But it is mentioned at commit 40e58a8a7ca6 ("dt-bindings: usb:
>>>>> Add binding for TI USB8041 hub controller").
>>>>>
>>>>> i2c-bus name property was suggested by Rob at
>>>>> https://lore.kernel.org/all/CAL_JsqJedhX6typpUKbnzV7CLK6UZVjq3CyG9iY_j5DLPqvVdw@xxxxxxxxxxxxxx/
>>>>> and
>>>>> https://lore.kernel.org/all/CAL_JsqJZBbu+UXqUNdZwg-uv0PAsNg55026PTwhKr5wQtxCjVQ@xxxxxxxxxxxxxx/
>>>>>
>>>>> the question is if adding address like this is acceptable.
>>>>> But it must be specified.
>>>>
>>>> Why? phandle points it explicitly.
>>>
>>> Ok it means just list usb hub on i2c with label and point to it. Works for me.
>>
>> Right. I missed you want the address of the hub but phandle goes to the
>> bus. I think listing it on I2C bus (see
>> arch/arm/boot/dts/vf610-zii-scu4-aib.dts) should work. I think we can
>> have I2C devices without compatibles.
>
> Device is definitely on i2c bus. But the problem with phande to bus is that
> there could more the same usb hubs and different i2c addresses of it. That's why
> I need to have exact match.
> Marek has similar hub where i2c address can be strapped too.
>
>> The problem is that property should have only one definition/type and
>> i2c-bus is already used in other cases as just "phandle". If we go with
>> your phandle+address approach, then this should be phandle-array with
>> items and then we have two different types.
>
> What to do with it then?

Your idea. I think you missed part of my comment. Add hub to the I2C bus
and phandle to the hub I2C device node.

Best regards,
Krzysztof