Commit 76c2ff3e broke backlight on Chromebook snow

[Simon Glass]

Hi,

On Sat, 23 Oct 2021 at 03:23, Dario Binacchi <dariobin at libero.it> wrote:
>
> Hi Guillaume,
>
> > Il 22/10/2021 14:34 Guillaume GARDET <guillaume.gardet at free.fr> ha scritto:
> >
> >
> > Hi Dario,
> >
> > ----- Dario Binacchi <dariobin at libero.it> a écrit :
> > > Hi Guillaume,
> > >
> > > > Il 21/10/2021 16:47 Guillaume GARDET <guillaume.gardet at free.fr> ha scritto:
> > > >
> > > >
> > > > Hi,
> > > >
> > > > Commit 76c2ff3e [0] broke backlight on Chromebook snow.
> > > > A revert on top of 2021.10 restores the backlight.
> > > > Not sure why it breaks on Chromebook snow yet. Dario, any idea how to fix this?
> > >
> > > Is it possible that max_level is equal to min_level and therefore division
> > > by zero is done?
> >
> > I do not think so, because u-boot does not crash and the DTS file has multiple levels:
> > https://source.denx.de/u-boot/u-boot/-/blob/master/arch/arm/dts/exynos5250-snow.dts#L243
> >
> > I think the problem occurs when 'cur_level == max_level', which is also the default value in the DTS.
> > If I change the default-brightness-level to 6 (instead of 7), then the backlight works again, which would confirm the assumption above.
>
> I have set the default-brightness-level to max-level on my beaglebone board but the backlight works the same.
> I didn't see any difference.

I got this running in my lab now and sent a possible fix.

Regards,
Simon

>
> Thanks and regards
> Dario
>
> > pwm_backlight_set_brightness seems to set cur_level to max value by default. Maybe the driver thinks there is nothing to do, since current level is also the desired level?
[..]