Created attachment 88022 [details] dmesg (its brightness can changed by the function keys) System Environment: -------------------------- Kernel: (drm-intel-next-queued)d538bbdfde34028b5c5b0ba92b3c2096c5afb82c Author: Damien Lespiau <damien.lespiau@intel.com> Date: Mon Oct 21 14:29:30 2013 +0100 Bug detailed description: ----------------------------- Everytime when the computer boot up into operating system, its eDP monitor will turn into black and its brightness is zero. But in the BIOS stage, the monitor works well. After the machine boot up, we can change the brightness with the function keys(Fn+F11),or we can also change the value of brightness in file (/sys/class/backlight/intel_backlight/brightness) .Then the monitor will work normally. But sometime the function keys doesn't work, and I can't find the file(/sys/class/backlight/intel_backlight/brightness). This occured once out of every two times. Reproduce steps: ----------------------------- 1.boot up the machine 2.cat /sys/class/backlight/intel_backlight/brightness
Created attachment 88023 [details] dmesg(its brightness can't be changed by the function keys) This dmesg is for the phenomenon when the brightness can't be changed by the function keys and also can't find the file (/sys/class/backlight/intel_backlight/brightness).
(In reply to comment #1) > This dmesg is for the phenomenon when the brightness can't be changed by the > function keys and also can't find the file > (/sys/class/backlight/intel_backlight/brightness). Very strange. The dmesg shows we get the brightness requests from the hotkeys, and they increase brightness but don't go very high. Did you try to press brightness up many times? Also, there's no indication why the sysfs interface would not show up. I don't understand. Can you try the patches from Aaron http://mid.gmane.org/1381498066-16011-1-git-send-email-aaron.lu@intel.com with video.use_native_backlight=1 module parameter? I believe that might fix most issues, and then we'd "only" have the vanishing sysfs to figure out...
Created attachment 88097 [details] dmesg of kernel with the patches In the test where I got dmesg in Comment 1 , I do have use hot keys to change the brightness , but the monitor didn't brighten up . I tried the patches you gived above with video.use_native_backlight=1 module parameter. But the bug still exit. I am afraid I didn't make it clear that the eDP monitor turn into black when it start to show the grub list. So it's already black before loading in the kernel. I hope this dmesg could be helpful.
(In reply to comment #3) > I am afraid I didn't make it clear that the eDP monitor turn into black when > it start to show the grub list. So it's already black before loading in the > kernel. So it shows the grub menu, and while showing the menu, goes black? Please also try video.use_bios_initial_backlight=0 parameter.
Yes It does show the grub menu, but it will go black when it start to load kernel after I choose one and click "enter". I tried video.use_bios_initial_backlight=0 parameter on latest drm-intel-next-queued branch. The eDP works well and won't turn into black again. And it seems that there is no difference whether I use the patch and "video.use_native_backlight=1" module parameter or not.
(In reply to comment #5) > I tried video.use_bios_initial_backlight=0 parameter on latest > drm-intel-next-queued branch. The eDP works well and won't turn into black > again. Aaron, possibly another one for an ACPI initial backlight quirk. Although I'm slightly surprised your win8 native backlight patches don't help. For future reference, would you like to have new bug reports on bugzilla.kernel.org for these, or is it okay to re-assign the bugs here to you?
(In reply to comment #6) > (In reply to comment #5) > > I tried video.use_bios_initial_backlight=0 parameter on latest > > drm-intel-next-queued branch. The eDP works well and won't turn into black > > again. > > Aaron, possibly another one for an ACPI initial backlight quirk. Although > I'm slightly surprised your win8 native backlight patches don't help. Me too, surprised. The video.use_bios_initial_backlight=0 will cause ACPI video not do a backlight level set based on value from _BQC, which might be buggy for a first time evaluation. But if my win8 native backlight patches got applied and video.use_native_backlight=1 is set, that set is also skipped, so it should solve the problem too... Hi Qingshuai, Can you please verify, with the win8 native backlight patch applied and video.use_native_backlight=1 set in kernel cmdline, there is no acpi_video0 in /sys/class/backlight. > > For future reference, would you like to have new bug reports on > bugzilla.kernel.org for these, or is it okay to re-assign the bugs here to > you? I'm OK with both ways :-)
For ACPI bugs, it is probably better to fill a new bug in kernel bugzilla.
Created attachment 88330 [details] dmesg with win8 patch and video.use_native_backlight=1 I tested the latest kernel on drm-intel-next-queued branch with the patch applied and video.use_native_backlight=1 set in kernel cmdline. And there IS the directory acpi_video0 in /sys/class/backlight. >For ACPI bugs, it is probably better to fill a new bug in kernel bugzilla. Do you mean that you want me to report this bug on bugzilla.kernel.org?
(In reply to comment #9) > Created attachment 88330 [details] > dmesg with win8 patch and video.use_native_backlight=1 > > I tested the latest kernel on drm-intel-next-queued branch with the patch > applied and video.use_native_backlight=1 set in kernel cmdline. > And there IS the directory acpi_video0 in /sys/class/backlight. Then this system's firmware must not claim win8 support. > > > >For ACPI bugs, it is probably better to fill a new bug in kernel bugzilla. > > Do you mean that you want me to report this bug on bugzilla.kernel.org? Yes please, and please attach acpidump when submitting this bug, thanks. Choose ACPI / Power-Video when submitting.
Sorry that I don't know how to get the acpidump. Can you please tell how to get it? Thanks. And when I get the acpidump, shall I use the kernel with the with the patch applied and video.use_native_backlight=1 set or without?
(In reply to comment #11) > Sorry that I don't know how to get the acpidump. Can you please tell how to > get it? Thanks. Run as root # acpidump > acpidump.txt The acpidump is available in the kernel tree: $kernel_tree/tools/power/acpi, just type make to get it. Or your distro may have packaged one. > And when I get the acpidump, shall I use the kernel with the with the patch > applied and video.use_native_backlight=1 set or without? It doesn't matter.
I have reported the bug on bugzilla.kernel.org. Its number is Bug 64031 .
(In reply to comment #13) > I have reported the bug on bugzilla.kernel.org. https://bugzilla.kernel.org/show_bug.cgi?id=64031 NOTOURBUG.
Closing old verified.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.