[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <4edab8d2-5b01-c406-332a-49a7305df5c0@gmail.com>
Date: Thu, 3 Aug 2023 19:19:38 +0200
From: Raphaël Gallais-Pou <rgallaispou@...il.com>
To: Conor Dooley <conor@...nel.org>,
Uwe Kleine-König <u.kleine-koenig@...gutronix.de>,
fabrice.gasnier@...s.st.com
Cc: Thierry Reding <thierry.reding@...il.com>,
Rob Herring <robh+dt@...nel.org>,
Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
Conor Dooley <conor+dt@...nel.org>,
Patrice Chotard <patrice.chotard@...s.st.com>,
linux-pwm@...r.kernel.org, devicetree@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH] dt-bindings: pwm: st: convert sti-pwm to DT schema
Hi,
Le 03/08/2023 à 18:09, Conor Dooley a écrit :
> On Thu, Aug 03, 2023 at 10:56:45AM +0200, Uwe Kleine-König wrote:
>> On Thu, Aug 03, 2023 at 09:18:14AM +0200, Raphaël Gallais-Pou wrote:
>>> Hi
>>>
>>> Le 02/08/2023 à 10:02, Uwe Kleine-König a écrit :
>>>> Hello,
>>>>
>>>> On Wed, Aug 02, 2023 at 12:05:59AM +0200, Raphael Gallais-Pou wrote:
>>>>> + st,capture-num-chan:
>>>>> + $ref: "/schemas/types.yaml#/definitions/uint32"
>>>>> + description: Number of available Capture channels.
>>>>
>>>> I have the theory that nobody actually uses the capture feature and I'd
>>>> like to get rid of it. People who do use it, should better switch to the
>>>> counter driver.
>>>
>>> TBH I only found two drivers using it, including this one.
>>>
>>> $ grep -rinI "\.capture" drivers/pwm/ | wc -l
>>> 2
>>
>> Right, there is pwm-stm32 and pwm-sti that support capture.
>>
>> There are a few machines that have a st,sti-pwm device:
>>
>> $ grep -rl st,sti-pwm arch/arm/boot/dts/*.dtb
>> arch/arm/boot/dts/stih407-b2120.dtb
>> arch/arm/boot/dts/stih410-b2120.dtb
>> arch/arm/boot/dts/stih410-b2260.dtb
>> arch/arm/boot/dts/stih418-b2199.dtb
>> arch/arm/boot/dts/stih418-b2264.dtb
>>
>> but to actually use capture the device tree must have a property
>> st,capture-num-chan. "st,capture-num-chan" isn't set by any of the
>> devices.
This is also what I came across, this is the reason why I'm not
reluctant to remove it.
>>
>> I think for stm32 it's not that trivial to show that it's unused.
>> While the capture code isn't a big maintenance burden, I still would
>> prefer to get rid of it if nobody uses it. Still more given that there
>> are better alternatives available.
Regarding stm32, I think the owner of the driver would prefer to handle it.
>>
>>> If there is no opposition about removing this feature I suggest to do it in
>>> a second time, in a serie.
>>
>> Does that mean you will do that? I guess not, but at least this means
>> you're not using capture support.
>
> It seems like it should either be done as part of the conversion or as a
> second patch in the series doing the conversion /shrug
Splitting the conversion and the capture removal is clearer IMO. Mixing
both could lead to confusion. I'll send another serie to do this.
Regards,
Raphaël
Powered by blists - more mailing lists