Re: [PATCH v3 09/20] ARM: shmobile: r8a7790: Add GPIO controller devices to device tree

[Laurent Pinchart <laurent.pinchart@ideasonboard.com>]

Hi Guennadi,

On Saturday 18 May 2013 08:57:41 Guennadi Liakhovetski wrote:
> On Sat, 18 May 2013, Laurent Pinchart wrote:
> > On Friday 17 May 2013 14:26:48 Guennadi Liakhovetski wrote:
> > > On Wed, 15 May 2013, Laurent Pinchart wrote:
> > > > Add GPIO controller nodes to the r8a7790 core device tree.
> > > > 
> > > > Signed-off-by: Laurent Pinchart 
> > > > <laurent.pinchart+renesas@ideasonboard.com>
> > > > ---
> > > > 
> > > >  arch/arm/boot/dts/r8a7790.dtsi | 54 +++++++++++++++++++++++++++++++++
> > > >  1 file changed, 54 insertions(+)
> > > 
> > > Also here a couple of things are missing (presumably, for other SoCs you
> > > need the same):
> >
> > Thank you for testing the patch set. I'll integrate the modification in
> > the next version for r8a7778, r8a7779 and r8a7790.
> > 
> > As the gpio-ranges property specifies the number of GPIOs handled by the
> > controller, what's your opinion on dropping the #gpio-lines property from
> > the bindings ?
> 
> Well, gpio-ranges can contain several ranges, so, in a generic case you
> would have to sum them up to get a total count. But if you're ok with that
> or if your platforms only have 1 gpio gange per controller, of course,
> that property seems redundant.

The gpio-rcar driver exposes a single range on all the current platforms. I 
don't really foresee any change there in the future, but I might be 
overlooking something. Magnus, any opinion ?

> OTOH it is already used in r8a7779.dtsi, and thus already belongs to the
> ABI... Not sure whether removing it would be accepted.

It's only used in my DT bindings proposal that haven't been merged yet, so 
that shouldn't be an issue :-)

-- 
Regards,

Laurent Pinchart