On Wed, Jun 26, 2013 at 7:35 PM, Petter Reinholdtsen <pere@xxxxxxxxxx> wrote: > [Daniel Vetter] >>> Is there some other quirk table for acpi that also need to be updated? >> >> If the acpi brightness controller is broken, it needs to be quirked >> away. For that please file a bug report against the ACPI subsystem. > > Right. Where do I do that? I can do it if I know what to do. :) > >> With that done the automagic fallback rules of the intel DDX driver >> should expose the "right" backlight interface through the backlight >> xrandr property. If KDE's powerdevil isn't using that one (and frobs >> around with the backlight controllers directly) that'd be a bug in >> kde. > > I discovered that by adding acpi_backlight=vendor and keeping the > i915.invert_brightness=1 setting, the screen went black again. So > these settings seem to cancel each other out. But if I remove > i915.invert_brightness=1, the screen no longer turn black, and KDE no > longer turn off the screen when I log in, because instead of using the > acpi_backlight interface it is using the intel_backlight interface, > which is not inverted. > > So I wonder, perhaps i915.invert_brightness=1 is the wrong quirk for > this machine? acpi_backlight=vendor should disable the acpi backlights (you can check that in /sys/class/backlight), which would mean that userspace should fall back to intel backlight driver. Can you please check whether that one works even without the invert_brightness knob? I.e. not just whether you'll end up with a black screen, but whether adjusting the backlight also works correctly. -Daniel -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch _______________________________________________ dri-devel mailing list dri-devel@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/dri-devel
