Re: [PATCH] dt-bindings: hwmon: Add nct7802 bindings

[Guenter Roeck]

On 9/10/21 1:44 PM, Oskar Senft wrote:
> Hi Guenter
>
> Thanks for the quick feedback!
>
> > > +  nuvoton,rtd-modes:
> > > +    description: |
> > > +      Select modes for the three RTDs.
> > > +
> > At the very least, "RTD" should be defined. The datasheet doesn't say explicitly,
> > but I suspect it means "Remote Temperature Diode".
> Ha, good point. As I understand, RTD means "Resistance Temperature
> Detector". But TBH, I'm not sure how that squares with Nuvoton's use
> of "LTD" for the local sensor ... sigh.
>
> > > +      Valid values for RTD1 and RTD2 are:
> > > +        "closed",
> > > +        "current",
> > > +        "thermistor",
> > > +        "voltage"
> > I am not sure what "closed" means (the datasheet doesn't say), but I suspect it means
> > that the sensor is disabled (?). For the other modes, the translation to the standard
> > ABI is:
> Thanks for that pointer, I now found that in
> Documentation/hwmon/sysfs-interface. Given that there's no definition
> for "disabled", I guess I'll just leave that out of the device tree
> binding for now? That way we'll stay consistent with the sysfs ABI.

Sure there is. A possible set of bindings - in that case for tmp421 -
is suggested with the series at
https://lore.kernel.org/linux-hwmon/cover.1631021349.git.krzysztof.adamski@xxxxxxxxx/

That specifically includes the ability to enable or disable channels
using the standard 'status' property. While that series is primarily
for the n-factor property supported by the tmp421, the same approach
can be used for [temperature] sensor properties on other chips as well.

I put [temperature] in [] because we'd need to find a means to express
if the sub-nodes are for temperature, voltage, or something else, but
I think the basic principle is sound.

Guenter

> That gives us the following mapping for sysfs / device tree -> nct7802 HW:
> 2 (3904 transistor) -> 3 (voltage)
> 3 (thermal diode) -> 1 (current)
> 4 (thermistor) -> 2 (thermistor)
>
> I'll update the device tree binding to be an array then. I also update
> the temp_type functions to support all 3 values.
>
> Oskar.