Re: [PATCH] arm64: dts: qcom: lemans-evk: Add host mode support for USB1 controller

[Swati Agarwal <swati.agarwal@oss.qualcomm.com>]

Hi,


On 12/3/2025 5:17 PM, Konrad Dybcio wrote:
> On 12/3/25 12:41 PM, Krishna Kurapati wrote:
>>
>>
>> On 12/3/2025 4:59 PM, Konrad Dybcio wrote:
>>> On 12/3/25 3:42 AM, Swati Agarwal wrote:
>>>> Enable Host mode support for USB1 controller on EVK Platform.
>>>>
>>>> For secondary USB Typec port, there is a genesys USB HUB sitting in between
>>>> SOC and HD3SS3220 Type-C port controller and SS lines run from the SoC
>>>> through the hub to the Port controller. Mark the second USB controller as
>>>> host only capable.
>>>>
>>>> Added HD3SS3220 Type-C port controller along with Type-c connector for
>>>> controlling vbus supply.
>>>>
>>>> Signed-off-by: Swati Agarwal <swati.agarwal@oss.qualcomm.com>
>>>> ---
> 
> [...]
> 
>>>> +                hd3ss3220_47_in_ep: endpoint {
>>>
>>> How about rename the other one to hd3ss3220_0 and name this one hd3ss322_1?
>>> _47 doesn't really tell us anything and may re-appear if there's another
>>> instance of this IC on another I2C bus

Thanks Konrad and Dmitry for the review.
For addressing this renaming for USB0 controller, shall i keep a separate patch and should i address in same DT patch for USB1?>>>
>>
>> ACK. Can we rename them as "usb-typec_1" and "usb_typec_2" ?
>> Krzysztof suggested to use generic names and hence we used "usb-typec" instead of hd3ss3220.
> 
> The generic names rule only applies to node names (text before '@'),
> the labels are generally only expected to "make overall sense", I think
> 
>>>> +&usb_1 {
>>>> +    dr_mode = "host";
>>>
>>> The connector states it's a dual-role data and power device. Is there any
>>> reason to keep this in host-only mode?
>>>
>>
>> As mentioned in commit text, there is a onboard hub sitting between SoC and the HD3SS3220 port controller. Hence device mode can't be used. This was the reason we left the above port nodes empty since we can't connect them to port nodes of controller.
> 
> It would have helped if I had paid more attention to that message then..
> 
>> Can we mark the connector as host only and remove the empty endpoints ? Would that we sufficient ?
> 
> The connector should definitely be marked host-only, but the endpoints should
> still reflect the physical connectivity.
> 
> If I understood your case properly, this is analogous to what &usb2_2_dwc3
> does in arch/arm64/boot/dts/qcom/sdm850-lenovo-yoga-c630.dts - see Commit
> c02716951e66 ("arm64: dts: qcom: sdm850-lenovo-yoga-c630: add routing for
> second USB connector")

For adding, onboard hub which is GL3590 in DT, it requires adding SS hub bindings in genesys,gl850g.yaml.
My plan is to first submit the binding patch and then follow up with this DT patch on top of it.

Regards,
Swati> 
> Konrad