Re: [PATCH 1/2] Documentation: dt-bindings: Use generic property for hardware enable pins

[Kim, Milo]

On 3/3/2017 3:21 PM, Rob Herring wrote:
> On Tue, Feb 28, 2017 at 04:50:40PM +0900, Milo Kim wrote:
> > With index usages, device specific properties can be replaced with generic
> > one. Vpos is index 0 and Vneg is index 1.
> > DT examples are added as well.
> >
> > Signed-off-by: Milo Kim <milo.kim@xxxxxx>
> > ---
> >  .../bindings/regulator/lm363x-regulator.txt        | 78 +++++++++++++++++++++-
> >  1 file changed, 76 insertions(+), 2 deletions(-)
> >
> > diff --git a/Documentation/devicetree/bindings/regulator/lm363x-regulator.txt b/Documentation/devicetree/bindings/regulator/lm363x-regulator.txt
> > index 8f14df9d1205..cc5a6151d85f 100644
> > --- a/Documentation/devicetree/bindings/regulator/lm363x-regulator.txt
> > +++ b/Documentation/devicetree/bindings/regulator/lm363x-regulator.txt
> > @@ -8,8 +8,8 @@ Required property:
> >
> >  Optional properties:
> >    LM3632 has external enable pins for two LDOs.
> > -  - ti,lcm-en1-gpio: A GPIO specifier for Vpos control pin.
> > -  - ti,lcm-en2-gpio: A GPIO specifier for Vneg control pin.
> > +  - enable-gpios: Two GPIO specifiers for Vpos and Vneg control pins.
> > +                  The first entry is Vpos, the second is Vneg enable pin.
>
> You're breaking compatibility with existing DTBs. You need to explain
> that and why it is okay in the commit message. In this case, I don't
> think it is okay as this chip could be used across vendors' platforms.

Thanks for your comment.

The lm363x-regulator has a dependency on ti-lmu MFD driver which is not 
upstreamed. So I don't think this patch will break the compatibility 
because two properties are not used anywhere at this moment.
Please correct me if it's incorrect.

Using general DT property is simple/clear because two enable pins are 
differentiable by selecting the index number. That's the main reason of 
this patch.

Best regards,
Milo