Re: [PATCH v2 1/4] dt-bindings: input: gpio-vibrator: Don't require enable-gpios

[OndÅej Jirman]

On Tue, May 12, 2020 at 03:52:12PM -0700, Dmitry Torokhov wrote:
> On Wed, May 13, 2020 at 12:22:02AM +0200, Ondrej Jirman wrote:
> > It is possible to turn the motor on/off just by enabling/disabling
> > the vcc-supply.
> > 
> > Signed-off-by: Ondrej Jirman <megous@xxxxxxxxxx>
> > Acked-by: Rob Herring <robh@xxxxxxxxxx>
> > ---
> >  Documentation/devicetree/bindings/input/gpio-vibrator.yaml | 1 -
> >  1 file changed, 1 deletion(-)
> > 
> > diff --git a/Documentation/devicetree/bindings/input/gpio-vibrator.yaml b/Documentation/devicetree/bindings/input/gpio-vibrator.yaml
> > index 2384465eaa19..c700b640bd53 100644
> > --- a/Documentation/devicetree/bindings/input/gpio-vibrator.yaml
> > +++ b/Documentation/devicetree/bindings/input/gpio-vibrator.yaml
> > @@ -24,7 +24,6 @@ properties:
> >  
> >  required:
> >    - compatible
> > -  - enable-gpios
> 
> Hmm we need at least one of the 2 (gpio and supply). Should we encode it
> in the binding?

Not sure how to encode either one or the other property being required, but
not both at once.

Maybe I can add a supply-vibrator compatible to the driver and binding and
make requirements dependent on the compatible?

> Also, in the dirver code, I guess we need to switch to have regulator
> optional (so we are not given the dummy one) and bail if neither
> regulator nor GPIO is found.

Though nothing bad will happen in the driver if binding will lack both
of these. The driver will just not control any HW, so at least it's
failsafe as is.

regards,
	o.

> Thanks.
> 
> -- 
> Dmitry