Re: [PATCH] dt-bindings: iio: afe: voltage-divider: Add io-channel-cells

[Peter Rosin]

Hi!

2024-01-26 at 17:16, Conor Dooley wrote:
> Hey,
> 
> On Fri, Jan 26, 2024 at 05:25:08PM +0530, Naresh Solanki wrote:
>> Add #io-channel-cells expected by driver. i.e., below is the message
>> seen in kernel log:
>> OF: /iio-hwmon: could not get #io-channel-cells for /voltage_divider1
>>
> 
>> TEST=Run below command & make sure there is no error:
>> make DT_CHECKER_FLAGS=-m dt_binding_check -j1
> 
> This shouldn't be in the commit message.
> 
>>
>> Signed-off-by: Naresh Solanki <naresh.solanki@xxxxxxxxxxxxx>
>> ---
>>  Documentation/devicetree/bindings/iio/afe/voltage-divider.yaml | 3 +++
>>  1 file changed, 3 insertions(+)
>>
>> diff --git a/Documentation/devicetree/bindings/iio/afe/voltage-divider.yaml b/Documentation/devicetree/bindings/iio/afe/voltage-divider.yaml
>> index dddf97b50549..b4b5489ad98e 100644
>> --- a/Documentation/devicetree/bindings/iio/afe/voltage-divider.yaml
>> +++ b/Documentation/devicetree/bindings/iio/afe/voltage-divider.yaml
>> @@ -39,6 +39,9 @@ properties:
>>      description: |
>>        Channel node of a voltage io-channel.
>>  
>> +  '#io-channel-cells':
>> +    const: 1
> 
> The example in this binding looks like the voltage-divider is intended
> to be an "IIO consumer" but "#io-channels-cells" is an "IIO provider"
> property.
> 
> Are you sure this is correct?

A voltage-divider is always an iio consumer. And like all iio things,
you may access its output from user space (typically via libiio). At
the same time a voltage-divider is optionally an iio provider for other
in-kernel thingies, in which case you need to specify
#io-channel-cells.

BTW, this is the case for for all bindings handled by the iio-rescale
driver.

Cheers,
Peter