Hi, On 05/15/2014 10:56 AM, Hans de Goede wrote: <snip> >>> So maybe we should simply drop the backlight_device_registered(raw) check? >> >> Unfortunately, there are indeed systems that with Intel GFX do not have >> a GPU backlight control interface: >> >> commit c675949ec58ca50d5a3ae3c757892f1560f6e896 >> Author: Jani Nikula <jani.nikula@xxxxxxxxx> >> Date: Wed Apr 9 11:31:37 2014 +0300 >> >> drm/i915: do not setup backlight if not available according to VBT >> >> And I remembered last time when we push the use_native default to 1 >> without checking if a raw interface is available, there are people >> complaining about no backlight interface is created on his system(and >> the only working interface is acpi_video on his Win8 system). So simply >> dropping this check doesn't seem like a good idea. > > Hmm, ok. So any smart ideas how to deal with the ordering problem we've > here ? > > Note this also plays into the proposal I'm about to send to unify and > simplify backlight control selection. Which besides just trying to > clean things up also tries to get rid of various module load ordering > issues. > > ... <this represent me thinking for half an hour trying to come up with a clever solution> > > So I think we really need some clean and generic way to deal with this, > which is not prone to module loading ordering issues, any suggestions? Ok, after yet more thinking about this I think I've what is likely going to be the best solution for this: 1) Add a callback to the backlight core which allows interested parties to get notified if a backlight device gets registered / unregistered 2) make acpi/video.c listen to these events and on these events re-check acpi_video_verify_backlight_support, and if it returns something different then the current situation (un)register the acpi_video# backlight devices This means that we will have ping-ponging of backlight interfaces which I really wanted to avoid, but given all the interdepencies that seems unavoidable. This will also simplify my cleanup proposal since if we accept the ping-ponging all the quirks can stay in the vendor specific firmware backlight control drivers. So, good or bad idea ? Regards, Hans -- To unsubscribe from this list: send the line "unsubscribe linux-acpi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
