Re: [PATCH V2 2/2] ARM: dts: bcm283x: Add CPU thermal zone with 1 trip point

[Eduardo Valentin <edubezval@gmail.com>]

Hey Stefan,


On Sun, Feb 19, 2017 at 01:31:49PM +0100, Stefan Wahren wrote:
> Hi Eduardo,
> 
> > Eduardo Valentin <edubezval@gmail.com> hat am 19. Februar 2017 um 02:17 geschrieben:
> > 
> > 
> > On Sun, Feb 12, 2017 at 07:38:48PM +0000, Stefan Wahren wrote:
> > > As suggested by Eduardo Valentin this adds the thermal zone for
> > > the bcm2835 SoC with its single thermal sensor. We start with
> > > the criticial trip point and leave the cooling devices empty
> > > since we don't have any at the moment.
> > > 
> > > Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
> > > ---
> > > 
> > > Changes in V2:
> > > - add missing thermal-sensor-cells property
> > > - change gpu-thermal to cpu-thermal
> > > 
> > >  arch/arm/boot/dts/bcm283x.dtsi |   20 ++++++++++++++++++++
> > >  1 file changed, 20 insertions(+)
> > > 
> > > diff --git a/arch/arm/boot/dts/bcm283x.dtsi b/arch/arm/boot/dts/bcm283x.dtsi
> > > index a3106aa..4dc74f6 100644
> > > --- a/arch/arm/boot/dts/bcm283x.dtsi
> > > +++ b/arch/arm/boot/dts/bcm283x.dtsi
> > > @@ -19,6 +19,25 @@
> > >  		bootargs = "earlyprintk console=ttyAMA0";
> > >  	};
> > >  
> > > +	thermal-zones {
> > > +		cpu_thermal: cpu-thermal {
> > > +			polling-delay-passive = <0>;
> > > +			polling-delay = <1000>;
> > > +
> > 
> > Check the diff I sent and also add the following for differentiating the
> > offsets and slopes depending on which chip the zone describes:
> > 			coefficients = <-538 407000>; /* for the zone on bcm2835 and bcm2836 */
> > 
> > and
> > 
> > 			coefficients = <-538 412000>; /* for the zone on bcm2837 */
> > 
> > 
> > Despite the changes mentioned for the driver and DT, I am ok with the driver and the DTS descriptors.
> 
> thanks for providing the necessary driver changes, but the coefficients above causes a DTC parse error. The Device Tree doesn't provide support for native signed integer. Looking at this old thread [1] suggests to add parentheses which fixed the parse issue. But of-thermal expected u32 for coefficients [2].
> 
> Any suggestions?

I am OK if you provide a patch to of-thermal in your series, assuming
that would fix the  representation issue the data of your driver has.

BR,

> 
> [1] - http://lists.infradead.org/pipermail/linux-arm-kernel/2013-April/159681.html
> [2] - http://elixir.free-electrons.com/source/drivers/thermal/of-thermal.c?v=4.10-rc7#L854
> 
> > 
> > BR,