Re: [PATCH 1/4] dt-bindings: mfd: qcom,spmi-pmic: add compatibles for pm4124-codec

[Krzysztof Kozlowski <krzk@kernel.org>]

On 11/02/2026 11:00, Srinivas Kandagatla wrote:
> On 2/11/26 6:43 AM, Krzysztof Kozlowski wrote:
>> On Mon, Feb 09, 2026 at 02:24:25PM +0000, Srinivas Kandagatla wrote:
>>> From: Alexey Klimov <alexey.klimov@linaro.org>
>>>
>>> Qualcomm Agatti SoC has PM4125 PMIC, which includes audio codec.
>>> Audio codec has TX and RX soundwire slave devices to connect to on-chip
>>> soundwire master.
>>>
>>> Add missing qcom,pm4125-codec compatible to pattern of audio-codec node
>>> properties in mfd qcom,spmi-pmic schema to complete the audio codec support.
>>>
>>
>> That's v6, not v1.
> Thanks Krzysztof, I was skeptical about it before sending however it
> made more sense to include that binding into this series which had 3 new
> patches,which is why I started of with v1 again. Which can be easy for
> review and dtb checks to not fail for the dts patches in this series.
> 
> Am happy to send it as v7 next time as a single patch if that is what is
> prefered.
> 
> 
> 
>>
>>> Signed-off-by: Alexey Klimov <alexey.klimov@linaro.org>
>>> [Srini: reworked the patch]
>>
>> Why? The v5 version was what we wanted. Why changing it?
>>
> I looked at all the versions of this patch before sending out this one,
> it sounded that you were not okay with the patch itself, may be that is
> not true, you just wanted a clear commit log changed. I can send a new
> version with updated comments as you requested. May be I miss
> interpreted it your review on v5.
> 
> TBH, there are multiple ways to add this new child node bindings into
> spmi bus.
> 
> Just from my understanding, why listing compatibles is preferred over
> listing ref to child bindings?, while both serve the same purpose.
> 

Both ways can be done here, but I proposed to keep it consistent, so
when adding a compatible, let's switch entire oneOf part (not entire
binding) to compatibles.

Later today I found conflicting patches, so maybe original approach is
actually easier - no conflicts, no rebases - let's go with it then:

Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>


Best regards,
Krzysztof