On Mon, 21 Jan 2019 10:49:17 +0100
Ibtsam Ul-Haq <ibtsam.haq.0x01@xxxxxxxxx> wrote:
> Hi,
>
> Thanks a lot for the review. I shall apply all these in the v2, I do
> have just one question, below.
>
> > > > +
> > > > +static int tla2024_of_find_chan(struct tla2024 *priv, struct device_node *ch)
> > > > +{
> > > > + u16 chan_idx = 0;
> > > > + u32 tmp_p, tmp_n;
> > > > + int ainp, ainn;
> > > > + int ret;
> > > > +
> > > > + ret = of_property_read_u32_index(ch, "single-ended", 0, &tmp_p);
> > There is now (only very recently) a generic description in DT for
> > acpi channels. Please look at that:
> >
> > https://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio.git/tree/Documentation/devicetree/bindings/iio/adc/adc.txt?h=togreg&id=00426e99789357dbff7e719a092ce36a3ce49d94
> >
> > It is minimalist. If it is single ended, you can tell that by the absence
> > of diff-channels.
> >
>
> How should we specify which single ended channel to use? For example,
> say I want to configure the driver for reading single-ended values
> only from input pin AIN3. When we only signify a single-ended channel
> by the absence of "diff-channels", where should the "3" be specified?
Fair question. Stefan, what do you think makes sense?
(obviously others welcome but my guess is Stefan has thought about this
more than most ;)
Jonathan
>
> > If the binding needs additional elements (it probably does)
> > then propose them for that generic binding. We are trying to move
> > away from each driver doing it's own thing given all this stuff is
> > pretty generic!
> >
>
> Sorry for trimming the message. Gmail was otherwise blocking my
> outbound message, and I could not find any other way to get it
> through.
>
> Best regards,
> Ibtsam Haq
