Re: [RFC PATCH v1 2/2] usb: host: add xhci-exynos to support Exynos SOCs

From: Krzysztof Kozlowski
Date: Mon Dec 05 2022 - 03:03:48 EST


On 05/12/2022 08:53, Jung Daehwan wrote:
> On Mon, Dec 05, 2022 at 08:33:39AM +0100, Krzysztof Kozlowski wrote:
>> On 05/12/2022 03:34, Jung Daehwan wrote:
>>
>>>>> Am I missing something here?
>>>>
>>>> Because it is not a driver for Exynos... it's a driver for wakelocks for
>>>> their specific Android use-cases which the manufacturer ships for their
>>>> Android devices. Due to Google GKI, they try to squeeze into upstream.
>>>> But this is huge misconception what should go to upstream and Samsung
>>>> does not want to keep discussing. They just send random patches and
>>>> disappear...
>>>>
>>>> Best regards,
>>>> Krzysztof
>>>>
>>>>
>>>
>>> No. It's driver for Exynos. Currently It only has wakelocks but I will
>>> submit one by one. Please think as the first patch of exynos not
>>> squeezed.
>>
>> That's not how upstream kernel development works... Your code has
>> nothing for Exynos. It's Android driver, not Exynos. If you say there is
>> something for Exynos it must be visible here. Wakelocks are not relevant
>> to Exynos, so after dropping them there would be empty stub in upstream
>> kernel which obviously cannot be accepted.
>>
>> Best regards,
>> Krzysztof
>>
>>
>
> Well, Exynos only uses wakelocks when I see mainline because it seems no

Exynos does not use wakelocks at all. Please explain me for what
hardware feature the wakelocks are needed when you do not use Android on
Exynos? Stop mixing Exynos with Android. One is hardware, second is
operating system.

> other driver use it. That's why I thought it could be a exynos specific.
> Do you agree that if I put wakelocks into xhci platform driver?

It's not related problem. Whether it suits there, I don't know.

Best regards,
Krzysztof