Re: [PATCH v3 4/5] dt-bindings: iio: adc: ad7192: Add AD7194 support

[Conor Dooley <conor@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 2786 bytes --]

On Thu, Feb 15, 2024 at 02:13:38PM +0200, Alisa-Dariana Roman wrote:
> On 08.02.2024 20:03, Conor Dooley wrote:
> > Hey,
> > 
> > On Thu, Feb 08, 2024 at 07:24:58PM +0200, Alisa-Dariana Roman wrote:
> > 
> > > +patternProperties:
> > > +  "^channel@([0-7a-f])$":
> > > +    type: object
> > > +    $ref: adc.yaml
> > > +    unevaluatedProperties: false
> > > +
> > > +    properties:
> > > +      reg:
> > > +        description: The channel index.
> > > +        minimum: 0
> > > +        maximum: 7
> > 
> > There are only 8 possible channels, at indices 0 to 7, so why is the
> > pattern property more permissive than that? Shouldn't "^channel@[0-7]$"
> > suffice?
> > 
> > > +
> > > +       diff-channels:
> > 
> > > +        description: |
> > > +          The differential channel pair for Ad7194 configurable channels. The
> > > +          first channel is the positive input, the second channel is the
> > > +          negative input.
> > 
> > This duplicates the description in adc.yaml
> > 
> > > +        items:
> > > +          minimum: 1
> > > +          maximum: 16
> > 
> > Hmm, this makes me wonder: why doesn't this match the number of channels
> > available and why is 0 not a valid channel for differential measurements?
> > 
> > Thanks,
> > Conor.
> 
> Hello and thank you for the feedback!
> 
> I will change the pattern property and the description.
> 
> Regarding the channels, I followed the existing style of the driver for the
> AD7194 channels: one iio channel for each pseudo-differential input
> channel(AINx - AINCOM), summing up to 16 channels; and one iio channel for
> each differential channel (AINx - AINy), summing up to 8 channels.

I don't know what question this is answering. Everything here is about
channels so it is hard for me to relate it back.
Please reply inline & not at the end of the message to everything :)
Was it meant to answer the following?

> > > +    properties:
> > > +      reg:
> > > +        description: The channel index.
> > > +        minimum: 0
> > > +        maximum: 7
> > 
> > There are only 8 possible channels, at indices 0 to 7, so why is the
> > pattern property more permissive than that? Shouldn't "^channel@[0-7]$"
> > suffice?

If it was a response to this, the reg property only allows 8 channels so
the regex should only allow 8 too. The number after @ must match the number
in reg. If using each of the 16 "pseudo-differential" inputs in
isolation is thing you want to be able to do, your reg property does not
allow it.

> For the
> diff-channels, I thought the possible values should be 1->16 corresponding
> to AIN1->AIN16 (I will add this to the description as suggested by David).

With a description, this should be fine.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]