Re: [PATCH v1 9/9] dt-bindings: iio: afe: add binding for temperature-sense-amplifier

[Jonathan Cameron <Jonathan.Cameron@Huawei.com>]

On Tue, 1 Jun 2021 17:02:51 +0100
Jonathan Cameron <Jonathan.Cameron@Huawei.com> wrote:

> On Mon, 31 May 2021 10:03:23 -0400
> "Liam Beguin" <liambeguin@gmail.com> wrote:
> 
> > Hi Peter,
> > 
> > On Mon May 31, 2021 at 3:32 AM EDT, Peter Rosin wrote:  
> > > Hi!
> > >
> > > On 2021-05-30 02:59, Liam Beguin wrote:    
> > > > From: Liam Beguin <lvb@xiphos.com>
> > > > 
> > > > An ADC is often used to measure other quantities indirectly. This
> > > > binding describe one cases, the measurement of a temperature through a
> > > > voltage sense amplifier such as the LTC2997.
> > > > 
> > > > Signed-off-by: Liam Beguin <lvb@xiphos.com>    
> > >
> > > What's the significant difference between this and the RTD binding? Does
> > > not both simply scale/offset a voltage to a temperature?  
> 
> I'm lost: what RTD binding?
Ignore this email - I was reading the series backwards and thought we were
talking about a preexisting binding.

> 
> > >    
> > 
> > The way I looked at it was one binding per sensor type (resistance
> > driven, current driven, and voltage driven).
> > 
> > Thinking about it more, these three bindings could be factorized into
> > one if the user is required to enter parameters "by hand".  
> 
> Don't. They are effectively different types of devices and we just end
> up with a more complex binding if we try to cover them all.
Ignore that as well. If the bindings combine fairly easily it is nice
to do so, but be careful not to throw too many things in together and
make it very hard to write the binding. However, I'm not keen on entirely
generic bindings and would like the channel type at least to come from
the compatible.

> 
> There is an argument to go the other way and actually have bindings for
> individual temperature sensors like the LTC2997.  Then the parameters
> become a driver problem rather than one for the binding.
> 
> Jonathan
> 
> 
> > 
> > These could become something like:
> > - sense-gain-mult
> > - sense-gain-div
> > - sense-offset
> > 
> > I like the idea of having the "datasheet parameters" in the devicetree,
> > but this would be a lot more versatile.
> > 
> > What do you think?
> > 
> > Cheers,
> > Liam
> >   
> > > Cheers,
> > > Peter    
> >   
>