Re: [PATCH v4 5/6] media: dt-bindings: ov5693: document YAML binding

[Daniel Scally]

Hello

On 30/06/2022 11:09, Tommaso Merciai wrote:
> Hi Sakari,
>
> On Thu, Jun 30, 2022 at 12:50:05PM +0300, Sakari Ailus wrote:
>> Hi Tommaso,
>>
>> On Thu, Jun 30, 2022 at 11:16:13AM +0200, Tommaso Merciai wrote:
>>> Hi Sakari,
>>>
>>> On Thu, Jun 30, 2022 at 12:12:47PM +0300, Sakari Ailus wrote:
>>>> On Thu, Jun 30, 2022 at 11:02:32AM +0200, Tommaso Merciai wrote:
>>>>> On Thu, Jun 30, 2022 at 10:07:19AM +0200, Krzysztof Kozlowski wrote:
>>>>>> On 30/06/2022 09:45, Tommaso Merciai wrote:
>>>>>>> Add documentation of device tree in YAML schema for the OV5693
>>>>>>> CMOS image sensor from Omnivision
>>>>>>>
>>>>>>> Signed-off-by: Tommaso Merciai <tommaso.merciai@xxxxxxxxxxxxxxxxxxxx>
>>>>>>> Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@xxxxxxxxxx>
>>>>>>> Reviewed-by: Sakari Ailus <sakari.ailus@xxxxxx>
>>>>>> How Sakari's tag appeared here? There was no email from him.
>>>>> Sakari made me some review on v2, but I think he forgot to add the mailing
>>>>> list in cc. ( I suppose :) )
>>>>>
>>>>> Let me know if I need to remove this.
>>>> You're only supposed to put these tags into patches if you get them in
>>>> written form as part of the review, signalling acceptance of the patch in
>>>> various forms. Just commenting a patch does not imply this.
>>>>
>>>> Please also see Documentation/process/submitting-patches.rst for more
>>>> information on how to use the tags.
>>> Thanks for sharing this. My bad.
>>> I remove your tags.
>> The patches themselves seem fine. I'd just drop the 4th patch or at least
>> come up with a better name for ov5693_hwcfg() --- you're acquiring
>> resources there, and that generally fits well for probe. The code is fine
>> already.
> Then we don't need v5 with your reviewed tags removed?
>
> I think the patch4 is needed to add dts support properly.
> Also this contains devm_clk_get_optional fix suggested by Jacopo and
> support for ACPI-based platforms that specify the clock frequency by
> using the "clock-frequency" property instead of specifying a clock
> provider reference.


I agree patch 4 in some form is needed - I didn't do the clock handling
particularly well in this driver, and though it's ostensibly an ACPI
driver it wouldn't actually work with a "normal" ACPI, but just with the
cio2-bridge-repaired style. So the changes to the clock handling logic
are welcome and needed I think. whether it needs to go into a separate
function I don't particularly mind either way.

>
> Some suggestion on the function name?
>
> Thanks,
> Tommaso
>
>> -- 
>> Sakari Ailus