Re: [PATCH v4 4/5] dt-bindings: iio: dac: Add adi,ltc2672.yaml

[Jonathan Cameron <jic23@kernel.org>]

On Sun, 23 Jun 2024 15:03:47 +0100
Conor Dooley <conor@kernel.org> wrote:

> On Sun, Jun 23, 2024 at 02:43:39PM +0100, Jonathan Cameron wrote:
> > On Wed, 19 Jun 2024 18:57:59 +0100
> > Conor Dooley <conor@kernel.org> wrote:
> >   
> > > On Wed, Jun 19, 2024 at 02:49:03PM +0800, Kim Seer Paller wrote:  
> > > > +patternProperties:
> > > > +  "^channel@[0-4]$":
> > > > +    type: object
> > > > +    additionalProperties: false
> > > > +
> > > > +    properties:
> > > > +      reg:
> > > > +        description: The channel number representing the DAC output channel.
> > > > +        maximum: 4
> > > > +
> > > > +      adi,toggle-mode:
> > > > +        description:
> > > > +          Set the channel as a toggle enabled channel. Toggle operation enables
> > > > +          fast switching of a DAC output between two different DAC codes without
> > > > +          any SPI transaction.
> > > > +        type: boolean
> > > > +
> > > > +      adi,output-range-microamp:
> > > > +        description: Specify the channel output full scale range.
> > > > +        enum: [3125000, 6250000, 12500000, 25000000, 50000000, 100000000,
> > > > +               200000000, 300000000]    
> > > 
> > > IIO folks, is this sort of thing common/likely to exist on other DACs?  
> > 
> > Fair point. It is probably time to conclude this is at least moderately common
> > and generalize it - which will need a dac.yaml similar to the one we have for
> > ADCs in adc/adc.yaml.  That will need to make this a per channel node
> > property (same as the adc ones).  
> 
> Looks like it is already per channel node?

Absolutely - but that is a bit fiddlier to do in a generic file so I just
meant make sure to enforce that in a similar fashion to adc.yaml.
> 
> > I'd also expect it to always take 2 values. In many cases the first will be 0
> > but that is fine.  
> 
> What would that first value represent?
Hmm. For voltage equivalent they are often not zero based so it would be the
negative. I have no idea if there are current dacs that act as both sources and sinks...
So maybe voltage one which should be done in this series as well needs to be 2 value
and current variant maybe not.

Jonathan