Re: [PATCH] nouveau: fix ambiguous backlight controls

[Jeremiah Mahler]

Ilia,

On Fri, Dec 26, 2014 at 04:39:08PM -0500, Ilia Mirkin wrote:
> On Fri, Dec 26, 2014 at 4:26 PM, Jeremiah Mahler <jmmahler@xxxxxxxxx> wrote:
> > If a display supports backlight control using the nouveau driver, and
> > also supports standard ACPI backlight control, there will be two sets of
> > controls.
> >
> > /sys/class/backlight/acpi_video0
> > /sys/class/backlight/nv_backlight
> >
> > This creates ambiguity because these controls can be out of sync with
> > each other.  One could be at 100% while the other is at 0% and the
> > actual display brightness depends on which one was used last.  This also
> > creates anomalies in Powertop which will show two values for brightness
> > with potentially different values.
> >
> > Fix this ambiguity by having the nouveau driver only enable its
> > backlight controls if the standard ACPI controls are not present.
> >
> > Signed-off-by: Jeremiah Mahler <jmmahler@xxxxxxxxx>
> > ---
> >  drivers/gpu/drm/nouveau/nouveau_backlight.c | 5 +++++
> >  1 file changed, 5 insertions(+)
> >
> > diff --git a/drivers/gpu/drm/nouveau/nouveau_backlight.c b/drivers/gpu/drm/nouveau/nouveau_backlight.c
> > index e566c5b..3a52bd4 100644
> > --- a/drivers/gpu/drm/nouveau/nouveau_backlight.c
> > +++ b/drivers/gpu/drm/nouveau/nouveau_backlight.c
> > @@ -221,6 +221,11 @@ nouveau_backlight_init(struct drm_device *dev)
> >         struct nvif_device *device = &drm->device;
> >         struct drm_connector *connector;
> >
> > +       if (acpi_video_backlight_support()) {
> 
> None of the other drivers have this. Is nouveau somehow different
> than, say, radeon in this respect?
> 
> Unfortunately the backlight situation is pretty fubar'd... sometimes
> the acpi controls don't work, sometimes the card controls don't work,
> sometimes they both work but in different ways (and then everyone's
> favourite -- neither works, and there's some third platform thing).
> I'm pretty sure this code existed before, but got removed. See commit
> bee564430feec1175ee64bcfd4913cacc519f817 and the previous commit
> 5bead799d3f8 before that. The ping-pong is probably not the right way
> to go.
> 

I was not aware of that change. But you are right, it took out what I
was trying to put back in.

Thanks for the helpful information.  I will have to rethink this fix.

[...]

-- 
- Jeremiah Mahler
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/