Re: [PATCH v4 2/2] power: reset: document LTC2952 poweroff support

[Mark Rutland <mark.rutland@arm.com>]

On Thu, Oct 09, 2014 at 05:09:36PM +0100, Guenter Roeck wrote:
> On Thu, Oct 09, 2014 at 04:07:56PM +0100, Mark Rutland wrote:
> > Hi,
> > 
> > On Thu, Oct 09, 2014 at 03:50:42PM +0100, Frans Klaver wrote:
> > > From: René Moll <linux@r-moll.nl>
> > > 
> > > Signed-off-by: René Moll <linux@r-moll.nl>
> > > Signed-off-by: Tjerk Hofmeijer <tjerk.hofmeijer@xsens.com>
> > > Signed-off-by: Frans Klaver <frans.klaver@xsens.com>
> > > ---
> > >  .../bindings/power/reset/ltc2952-poweroff.txt      | 31 ++++++++++++++++++++++
> > >  1 file changed, 31 insertions(+)
> > >  create mode 100644 Documentation/devicetree/bindings/power/reset/ltc2952-poweroff.txt
> > > 
> > > diff --git a/Documentation/devicetree/bindings/power/reset/ltc2952-poweroff.txt b/Documentation/devicetree/bindings/power/reset/ltc2952-poweroff.txt
> > > new file mode 100644
> > > index 0000000..c3f3d9e
> > > --- /dev/null
> > > +++ b/Documentation/devicetree/bindings/power/reset/ltc2952-poweroff.txt
> > > @@ -0,0 +1,31 @@
> > > +Binding for the LTC2952 PowerPath controller
> > > +
> > > +This chip is used to externally trigger a system shut down. Once the trigger
> > > +has been sent, the chips watchdog has to be reset to gracefully shut down.
> > 
> > s/chips/chip's/
> > 
> > > If +the Linux systems decides to shut down, it powers off the platform
> > > via the +poweroff signal.
> > 
> > This sentence can go; the binding should describe the hardware rather
> > than the Linux behaviour.
> > 
> Maybe find an operating system independent wording and describe what the
> implementation should do. Or is that out of scope as well ?

Something like "The platform can be powered off via the poweroff signal"
would be OK. So long as it's not a description of Linux internals that's
fine. 

Ideally the binding wouldn't state how the OS should use the device
either. So if bindings can be worded w.r.t. possible uses of a device
rather than specific users, that would be preferable unless the entire
point of the binding is to specify a very specific use.

Thanks,
Mark.