Hi Karol Babioch, On Fri, Aug 22, 2014 at 7:35 AM, Karol Babioch <karol@xxxxxxxxxx> wrote: > By looking at your description (and your patch) I'm wondering how > exactly this problem made itself apparent? I'm having a similar (the > same?) issue with my Sony Vaio VPCS12C5E. The backlight interface > registered by ACPI does not work and I have to boot with the command > line option "acpi_backlight=vendor". Even then two interfaces get > registered "nv_backlight" and "sony", only one of which (nv_backlight) > works. > My situation is that the brightness control on Lenovo B470e is broken on Ubuntu 12.04.5 but 12.04.2. Neither "acpi_backlight=vendor " nor "video.use_native_backlight=1" works for the brightness control. After bisecting, the regression happens in 3.10-rc1. Before 3.10-rc1, acpi/video.c fails to register backlight interface due to acpi_video_device_lcd_set_level() returns -EINVAL in acpi_video_init_brightness(). Only "intel_backlight" is registered under /sys/class/backlight. So the brightness can be adjusted from UI. In 3.10-rc1, some changes make acpi/video.c able to register backlight for Lenovo B470e. "acpi_video0" is used as the default backlight interface. There is no BIOS update since 2012. And the vendor is unlikely to fix this bug. So I propose the patch for fixing the brightness control on Lenovo B470e. > Desktop environments (GNOME in particular) seem to be confused by this, > so I always have to change the brightness by writing directly into the > "actual_brightness" file. I've reported this back in 2012 [1], but was > told that it is an ACPI issue and so it never was fixed. > Similar situation on Toshiba T130-15T https://bugzilla.kernel.org/show_bug.cgi?id=50551 > Couldn't a similar patch be applied to the sony-laptop module? I've > added Mattia Dongili to the discussion, hopefully he doesn't mind to > take a look at this again. I'm glad to test any patches, but am not > familiar enough with all of the internal structs to mess around with > them for myself without breaking support for other users of this module, > especially since my model was shipped with in two different versions: > One with the internal Intel GPU and another with a dedicated Nvidia GPU. > I'm afraid a simple DMI match might not be good enough in this case? > You may use dmidecode to see if there is any extra information which can differentiate between these two versions. Regards, Edward Lin -- To unsubscribe from this list: send the line "unsubscribe platform-driver-x86" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
