Re: [PATCH v4 2/5] dt-bindings: thermal: generic-adc: Document #io-channel-cells

From: Luca Weiss

Date: Thu Nov 19 2026 - 04:13:16 EST


On Wed Nov 18, 2026 at 5:42 PM CET, Rob Herring wrote:
> On Tue, Nov 17, 2026 at 09:23:12AM +0100, Luca Weiss wrote:
>> Document the property to allow using this sensor with io-channels.
>>
>> This means that the output of generic-adc-thermal can be used as an
>> input into a different node.
>>
>> As a specific example, generic-adc-thermal takes in an ADC io-channel in
>> Volt, outputs temperature in degC, which can then be consumed by a fuel
>> gauge driver to provide the battery temperature readings to user space.
>
> But this is already a thermal-sensor provider, why wouldn't the fuel
> gauge use that provider?

thermal-sensors = <&my_generic_adc_thermal>; could work, yes, but IIO is
way more flexible and can support the required use case very well.

>From what I can see no device binding (or Linux) uses thermal-sensors
apart from thermal zones, so actually doing this for another device
feels like re-inventing the wheel when IIO already exists.

Please also see the cover letter from the RFC[0] under "1. Do not
implement IIO device in generic-adc-thermal"

[0] https://lore.kernel.org/linux-arm-msm/20251010-bat-temp-adc-v1-0-d51ec895dac6@xxxxxxxxxxxxx/

Regards
Luca

> Seems like this is letting current Linux driver design dictate the
> binding.
>
> Rob