[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20120907201017.GA17043@avionic-0098.mockup.avionic-design.de>
Date: Fri, 7 Sep 2012 22:10:17 +0200
From: Thierry Reding <thierry.reding@...onic-design.de>
To: Sascha Hauer <s.hauer@...gutronix.de>
Cc: Shawn Guo <shawn.guo@...aro.org>,
linux-arm-kernel@...ts.infradead.org,
HACHIMI Samir <shachimi@...neo-embedded.com>,
linux-kernel@...r.kernel.org,
Benoît Thébaudeau
<benoit.thebaudeau@...ansee.com>
Subject: Re: [PATCH 8/9] ARM i.MX53: Add pwms to dtsi
On Fri, Sep 07, 2012 at 07:26:12PM +0200, Sascha Hauer wrote:
> On Fri, Sep 07, 2012 at 03:29:55PM +0200, Thierry Reding wrote:
> > On Fri, Aug 31, 2012 at 03:07:23PM +0200, Sascha Hauer wrote:
> > > On Fri, Aug 31, 2012 at 06:32:20AM +0800, Shawn Guo wrote:
> > > > On Tue, Aug 28, 2012 at 01:48:23PM +0200, Sascha Hauer wrote:
> > [...]
> > > > > arch/arm/boot/dts/imx53.dtsi | 14 ++++++++++++++
> > > > > arch/arm/mach-imx/clk-imx51-imx53.c | 4 ++++
> > > > > 2 files changed, 18 insertions(+)
> > > > >
> > > > > diff --git a/arch/arm/boot/dts/imx53.dtsi b/arch/arm/boot/dts/imx53.dtsi
> > > > > index cd37165..7ec17e4 100644
> > > > > --- a/arch/arm/boot/dts/imx53.dtsi
> > > > > +++ b/arch/arm/boot/dts/imx53.dtsi
> > > > > @@ -189,6 +189,20 @@
> > > > > status = "disabled";
> > > > > };
> > > > >
> > > > > + pwm1: pwm@...b4000 {
> > > > > + #pwm-cells = <3>;
> > > >
> > > > pwm-cells should be 2?
> > >
> > > Yes, right. We have a patch internally that allows us to pass a
> > > 'inverted' flag to the pwm, hence I accidently have 3 here.
> >
> > There are patches in for-next that add support for setting the PWM
> > polarity, though there's currently no support for specifying it via a
> > third cell in the specifier. Would you mind sharing the patches that add
> > this?
>
> Yes, will do. I was afraid this leads to some discussion, so I skipped
> them so far.
>
> The basic idea was that the third cell is for flags from which bit0 set
> means 'inverted'. We currently implemented this i.MX specific, but if
> you think this is acceptable it's propably a good idea to implement this
> in a generic manner.
Absolutely. That's exactly what I had in mind as well, so if you have
the code ready I don't have to write it myself.
Thierry
Content of type "application/pgp-signature" skipped
Powered by blists - more mailing lists