Re: [PATCH 1/2] dt-bindings: pwm-backlight: Add pwm-delay-us property

[Daniel Thompson]

On 28/06/17 15:15, Enric Balletbo Serra wrote:
> Hi Daniel, Pavel,
>
> 2017-06-28 15:30 GMT+02:00 Pavel Machek <pavel@xxxxxx>:
> > On Wed 2017-06-28 13:16:31, Enric Balletbo i Serra wrote:
> > > From: huang lin <hl@xxxxxxxxxxxxxx>
> > >
> > > Add a pwm-delay-us property to specify the delay between setting an
> > > initial (non-zero) PWM value and enabling the backlight, and also the
> > > delay between disabling the backlight and setting PWM value to 0.
> >
> >
> >
> > > Signed-off-by: huang lin <hl@xxxxxxxxxxxxxx>
> > > Signed-off-by: Enric Balletbo i Serra <enric.balletbo@xxxxxxxxxxxxx>
> > > ---
> > >   Documentation/devicetree/bindings/leds/backlight/pwm-backlight.txt | 4 ++++
> > >   1 file changed, 4 insertions(+)
> > >
> > > diff --git a/Documentation/devicetree/bindings/leds/backlight/pwm-backlight.txt b/Documentation/devicetree/bindings/leds/backlight/pwm-backlight.txt
> > > index 764db86..f75b08f 100644
> > > --- a/Documentation/devicetree/bindings/leds/backlight/pwm-backlight.txt
> > > +++ b/Documentation/devicetree/bindings/leds/backlight/pwm-backlight.txt
> > > @@ -17,6 +17,9 @@ Optional properties:
> > >                  "pwms" property (see PWM binding[0])
> > >     - enable-gpios: contains a single GPIO specifier for the GPIO which enables
> > >                     and disables the backlight (see GPIO binding[1])
> > > +  - pwm-delay-us: delay between setting an initial (non-zero) PWM value and
> > > +                  enabling the backlight, and also the delay between disabling
> > > +                  the backlight and setting PWM value to 0.
> >
> > > If is safe to assume power-on delay and power-off delay will be the same?
> > >
> > > I've only took a quick look but several backlight controllers support asymetric power-on/off sequences..
>
> Daniel, right the ones I checked are symmetric but asymmetric timings
> are possible I guess, so I think now that specify the two delays is
> more accurate, maybe the binding should be something like this?
>
>    pwm-delay-us = <delay_before_on  delay_after_off>;

I think so. Like you I can't actually point at any asymmetric power 
sequence diagrams but there are a controller devices with power 
sequencing registers that support asymmetry. I'm inclined to take that 
as a hint...


> > I understood it as "you set PWM and it takes a while for display to
> > light up"
> > but that's not correct. Changelog from second patch makes it
> > clear. Please
> > clarify it here, too.
>
> Pavel, oh, your dude is interesting ...
>
> That's the idea, the sequence is:
>    Power on, you set the PWM signal, wait a bit and set the LED_EN signal.
>    Power off, you clear the LED_EN signal, wait a bit and stop the PWM signal.
>
> Note that the patch inverts the sequence, before this patch first you
> set LED_EN signal and then the PWM signal
>
> I assumed that the sequence was wrong but maybe I'm mistaken and there
> are some panels that follow the original sequence. On the few panels I
> checked the power on/off sequence is how I described above, i.e. see
> [1] p. 17, the sequence is first PWM and then LED_EN. I'll take a look
> at other panel datasheets, or if you know one, could you provide me
> the datasheet?
>
> Thanks,
>
> [1] http://www.jxlcd.com/Upload/PicFiles/N116BGE-L41.pdf
>
> >                                                                  Pavel
> > --
> > (english) http://www.livejournal.com/~pavelmachek
> > (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html