[PATCH] intel_display.c: make acpi_video last in backlight_interfaces

Chris Wilson chris at chris-wilson.co.uk
Sun Aug 26 11:57:14 PDT 2012

On Sun, 26 Aug 2012 16:52:44 +0200, Alon Levy <alevy at redhat.com> wrote:
> On Sun, Aug 26, 2012 at 03:45:15PM +0100, Chris Wilson wrote:
> > On Sun, 26 Aug 2012 16:31:41 +0200, Alon Levy <alevy at redhat.com> wrote:
> > > OK. I'm not sure what would be better for gnome / your-favorite desktop
> > > manager. One option is to write an xorg.conf file with this patch.
> > > Another is a property for the backlight interface. Even though it's
> > > abuse of the property interface for a one time thing (not like it would
> > > change during the lifetime of X). Right now it's a stopgap thing, but
> > > not very user friendly. This patch came from my brother complaining that
> > > with his mac he can lower the backlight all the way to 0, but not with
> > > his linux laptop. I don't want to have to maintain an xorg.conf file.
> > > 
> > > Sorry for the blurb - bottom line, I'm not sure if an xorg.conf file
> > > would be something management, i.e. libgnome-desktop / kde etc. would
> > > want to deal with, setting a property is perhaps easier, or another
> > > solution.
> > 
> > I would argue that the manufacturer expected the OS to use DPMS to turn
> > the backlight completely off and save further power when opting to have
> > it minimal backlight setting at non-zero.
> > 
> That sounds right, thanks! Now it just becomes something that the
> application (in my case libgnomedesktop) has to deal with, i.e. if
> laptop brand X use DPMS for 0 brightness, else not. Or maybe always
> using it would also work. I guess this patch is not required in this
> case, better drop it unless you know, you like it :)

It helps with another problem we face where we can choose the wrong
interface in a multi-panel system, so I've applied a variation as a step
along that path. Thanks, I hope you find a satisfactory solution to your
power-saving quest :)

Chris Wilson, Intel Open Source Technology Centre

More information about the xorg-devel mailing list