Re: Driver directory selection for Power Supply Status chip

[Guenter Roeck <linux@roeck-us.net>]

On Mon, Apr 18, 2016 at 09:06:02PM +0100, Jonathan Cameron wrote:
> On 18/04/16 20:43, Thor Thayer wrote:
> > 
> > 
> > On 04/18/2016 11:41 AM, Lars-Peter Clausen wrote:
> >> On 04/18/2016 06:14 PM, Thor Thayer wrote:
> >>> Hi,
> >>>
> >>> The Arria10 System Resource chip (A10SR) is a SPI based MFD implementing a
> >>> GPIO expander, reset controller, and power supply monitoring. I'm not sure
> >>> where the driver for the A10SR power monitoring driver should reside and I'm
> >>> hoping someone can offer some advice.
> >>>
> >>> I'd originally submitted the RFC to the HWMON list but the maintainer
> >>> pointed out that it wasn't a good fit because the A10SR only indicates
> >>> boolean power supply status - not the voltage level as required by HWMON.
> >>>
> >>> I read that IIO acts as a bridge between IO and HWMON but I'm also not sure
> >>> this fits those drivers (although I did find some power supply supervistors
> >>> there). It isn't quite a power supply supervisor - the A10SR is a comparator
> >>> instead of an Analog-to-Digital Converter.
> >>>
> >>> One additional thing, the A10SR also had a number of enable bits for
> >>> enabling devices external to Altera's SoC (but still on the development
> >>> board). These don't quite fit into the reset controller framework but would
> >>> seem to fit a MISC directory driver better.
> >>>
> >>> If neither IIO or MISC is a good fit, can someone suggest a more appropriate
> >>> place?
> >>
> >> How does the device work, does it generate an interrupt when the voltage
> >> level goes below the threshold? I'd like to pass the ball back and say that
> >> this sounds like something that should go into hwmon.
> >>
> > 
> > Good question but no, there is no interrupt.
> > 
> > The chip is actually a power supply sequencer for bringing up the
> > power supplies in the correct order. Since the sequence, timings, and
> > thresholds are hard coded in the chip and can't be changed
> > programatically, I assume it was decided that a power fail interrupt
> > was not needed. However, the status of the power (OK/Fail) can be
> > read from the chip.
> > 
> > I was directed to look at Documentation/hwmon/sysfs-interface which
> > states the units are in millivolts and I didn't see any references to
> > a boolean output so I'm leaning away from hwmon.
> That's one impressively uninformative output...  Hmm.

Please clarify what you mean with "impressively uninformative output".

FWIW, hwmon alarm attributes are boolean.

Thanks,
Guenter

> Not obvious where to put it - it doesn't fit in IIO really either.
> 
> Could report as a power supply?  There is 'health' support in there.
> See Documentation/power/power_supply_class.txt
> and include/linux/power_supply.h
> though I'm not sure which type a fail on this would count as...
> perhaps UNSPEC_FAILURE.
> 
> It's intended for batteries really - not sure how far you can stretch
> that.
> 
> Cc'd linux-pm and maintainers.  Perhaps they will have a better idea!
> > 
> > Thank you for your reply!
> > -- 
> > To unsubscribe from this list: send the line "unsubscribe linux-iio" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
>