[IGC Mobile 4] Screen brightness not 100% at boot or resume from monitor sleep

[Charles Profitt]

Manufacturer: LENOVO
Version: ThinkPad T500

This happened with the 2.6.27-11-generic kernel update on January 28th. Since this update when the laptop boots the screen is dim (using the keyboard controls raises the brightness to 100%) or if the monitor goes to sleep and I wake the monitor up the same issue happens.

I ran the gnome-power-bugreport.sh script and will attach the out put of that as well as gnome-power-manager --no-daemon -- verbose

00:02.0 VGA compatible controller [0300]: Intel Corporation Mobile 4 Series Chipset Integrated Graphics Controller [8086:2a42] (rev 07)
00:02.1 Display controller [0380]: Intel Corporation Mobile 4 Series Chipset Integrated Graphics Controller [8086:2a43] (rev 07)

[Charles Profitt (cprofitt)]

Here is the output of gnome-power-manager --no-daemon --verbose 2 >&1 tee gpm.debug.log.txt

TI:23:46:02 TH:0x870a00 FI:gpm-brightness-xrandr.c FN:gpm_brightness_xrandr_output_get_percentage,225
 - percentage 86
TI:23:46:02 TH:0x870a00 FI:gpm-feedback-widget.c FN:gpm_feedback_display_value,140
 - Displaying 0.860000 on feedback widget
TI:23:46:02 TH:0x870a00 FI:gpm-refcount.c FN:gpm_refcount_add,100
 - refcount now: 6
TI:23:46:02 TH:0x870a00 FI:gpm-refcount.c FN:gpm_refcount_add,101
 - non zero, so sending REFCOUNT_ADDED
TI:23:46:02 TH:0x870a00 FI:gpm-backlight.c FN:gpm_backlight_button_pressed_cb,567
 - emitting brightness-changed : 86
TI:23:46:02 TH:0x870a00 FI:gpm-info.c FN:button_pressed_cb,679
 - Button press event type=brightness-down
TI:23:46:02 TH:0x870a00 FI:gpm-brightness-xrandr.c FN:gpm_brightness_xrandr_filter_xevents,552
 - Event0 0x9d0280
TI:23:46:03 TH:0x870a00 FI:gpm-button.c FN:hal_device_condition_cb,430
 - condition=ButtonPressed, details=brightness-down
TI:23:46:03 TH:0x870a00 FI:gpm-button.c FN:emit_button_pressed,374
 - emitting button-pressed : brightness-down
TI:23:46:03 TH:0x870a00 FI:gpm-manager.c FN:button_pressed_cb,1020
 - Button press event type=brightness-down
TI:23:46:03 TH:0x870a00 FI:gpm-srv-screensaver.c FN:button_pressed_cb,167
 - Button press event type=brightness-down
TI:23:46:03 TH:0x870a00 FI:gpm-backlight.c FN:gpm_backlight_button_pressed_cb,536
 - Button press event type=brightness-down
TI:23:46:03 TH:0x870a00 FI:gpm-brightness-xrandr.c FN:gpm_brightness_xrandr_foreach_screen,414
 - using resource 0x9e3e00
TI:23:46:03 TH:0x870a00 FI:gpm-brightness-xrandr.c FN:gpm_brightness_xrandr_foreach_resource,375
 - resource 1 of 2
TI:23:46:03 TH:0x870a00 FI:gpm-brightness-xrandr.c FN:gpm_brightness_xrandr_foreach_resource,375
 - resource 2 of 2
TI:23:46:03 TH:0x870a00 FI:gpm-brightness-xrandr.c FN:gpm_brightness_xrandr_output_down,250
 - hard value=13, min=0, max=15
TI:23:46:03 TH:0x870a00 FI:gpm-brightness-xrandr.c FN:gpm_brightness_xrandr_foreach_screen,414
 - using resource 0x9e3e00
TI:23:46:03 TH:0x870a00 FI:gpm-brightness-xrandr.c FN:gpm_brightness_xrandr_foreach_resource,375
 - resource 1 of 2
TI:23:46:03 TH:0x870a00 FI:gpm-brightness-xrandr.c FN:gpm_brightness_xrandr_foreach_resource,375
 - resource 2 of 2
TI:23:46:03 TH:0x870a00 FI:gpm-brightness-xrandr.c FN:gpm_brightness_xrandr_output_get_percentage,223
 - hard value=12, min=0, max=15
TI:23:46:03 TH:0x870a00 FI:gpm-brightness-xrandr.c FN:gpm_brightness_xrandr_output_get_percentage,225
 - percentage 80
TI:23:46:03 TH:0x870a00 FI:gpm-feedback-widget.c FN:gpm_feedback_display_value,140
 - Displaying 0.800000 on feedback widget
TI:23:46:03 TH:0x870a00 FI:gpm-refcount.c FN:gpm_refcount_add,100
 - refcount now: 7
TI:23:46:03 TH:0x870a00 FI:gpm-refcount.c FN:gpm_refcount_add,101
 - non zero, so sending REFCOUNT_ADDED
TI:23:46:03 TH:0x870a00 FI:gpm-backlight.c FN:gpm_backlight_button_pressed_cb,567
 - emitting brightness-changed : 80
TI:23:46:03 TH:0x870a00 FI:gpm-info.c FN:button_pressed_cb,679
 - Button press event type=brightness-down
TI:23:46:03 TH:0x870a00 FI:gpm-button.c FN:gpm_button_filter_x_events,122
 - Key 232 mapped to HAL key brightness-down
TI:23...

[Matthew Lye (matthew.lye)]

Tested the same on my HP Compaq nc6400

[Connor Imes (ckimes)]

Thank you for reporting this Charles, I have a lot of information to request since this problem could have a number of potential causes.

Can you please include the information as separate attachments:
   cat /proc/version_signature > proc_version_signature.txt
   lsb_release -rd > lsb_release.txt
   uname -a > uname.txt
   sudo lspci -vvnn > lspci.txt
   sudo dmidecode > dmidecode.txt
   lshal > lshal.txt
   gconftool --recursive-list /apps/gnome-power-manager > gpm.gconf.values.txt
and
   /var/log/kern.log.0
as well as tarball of /proc/acpi directory. You can't just tar all files because their content sometimes changes etc.
    * cp -r /proc/acpi /tmp
    * tar -cvjf ~/acpi.tar.bz /tmp/acpi
    * attach acpi.tar.bz from your home directory

Thanks in advance.

[Charles Profitt (cprofitt)]

I was unable to complete the last request due to this

$ cp -r /proc/acpi /tmp
cp: cannot open `/proc/acpi/event' for reading: Permission denied
cp: cannot open `/proc/acpi/fadt' for reading: Permission denied
cp: cannot open `/proc/acpi/dsdt' for reading: Permission denied
$ sudo cp -r /proc/acpi /tmp
cp: cannot open `/proc/acpi/event' for reading: Device or resource busy

[Charles Profitt (cprofitt)]

2.6.27-9-generic -- kern.log

[Charles Profitt (cprofitt)]

The lshal from the 2.6.27-9 kernel

[Connor Imes (ckimes)]

Thanks for providing the extra information I requested on IRC.

The lshal from the older kernel shows
   udi = '/org/freedesktop/Hal/devices/computer_backlight_0'
   udi = '/org/freedesktop/Hal/devices/computer_backlight'
whereas the newer kernel only shows the latter.

Your gconf settings look OK to me, and the kernel logs seem consistent (at least to my relatively untrained eye).

Because this appears to be a hardware detection problem, I'm changing this bug to the kernel.

I guess I forgot to ask for dmesg, can you please also attach
   dmesg > dmesg.txt
from both kernels since you're currently in the old one anyway. Again, please be sure to describe the one from the older kernel.

Thanks.

[Connor Imes (ckimes)]

This bug is currently filed under "linux" because of hardware behavior regression, though it may need to be re-assigned to the intel graphics driver (xserver-xorg-video-intel) or an acpi package (e.g. acpi-support).

According to term.log, the version of the intel driver is 2:2.4.1-1ubuntu10.3 which is the latest in Intrepid, and was updated on 1/23/2009 - only shortly prior to the kernel upgrade.
Charles, can you please attach (from inside the latest kernel)
   /var/log/gdm/\:0.log
   /var/log/Xorg.0.log

Thanks again.

[Connor Imes (ckimes)]

Charles,
Can you please try adding the following entry to /etc/modprobe.d/options (without quotes):
   "options thinkpad_acpi backlight=1"
Then reboot into the latest kernel and see if the problem persists. Also please post your kern.log after doing that, and let us know the outcome.
Thanks.

[Charles Profitt (cprofitt)]

It did not do anything.

[Charles Profitt (cprofitt)]

The behavior is not on my T61p which has an nvidia card nor on my T42p which uses an ATI card. Neither of the other two laptops are Centrino 2

[Connor Imes (ckimes)]

Ok, thanks Charles. It seems that it didn't like that option that was added to /etc/modprobe.d/options - you can remove it now, it was a suggestion from some one in #ubuntu-kernel.
With it, we see
   Jan 30 19:50:15 Tardis-T500 kernel: [ 12.263200] thinkpad_acpi: Unknown parameter `backlight'

I'll mark this bug as Triaged now so a developer can look at it. There is a wealth of information here, though they might ask for more. Thank you again for reporting this.

[Charles Profitt (cprofitt)]

Is there anyway to know if a developer has looked at this or anything more I can do?

[Dominik Stadler (dominik-stadler)]

As a workaround I use the following script as /etc/pm/sleep.d/01LCD to fix this for resume from sleep/hibernate. This is on a Thinkpad T500. A similar thing should be possible somewhere in the boot scripts as well.

#!/bin/sh
# Workaround for LCD brightness bug
# Save this as /etc/pm/sleep.d/01LCD

#. "${PM_FUNCTIONS}"
. "/usr/lib/pm-utils/functions"

case "$1" in
        hibernate|suspend)
                ;;
        thaw|resume)
                # for some reason the backlight brightness is reset and not set correctly any more, try
                # to get and set the current brightness to re-set the correct one here

                BACKLIGHT=`cat /proc/acpi/video/VID/LCD0/brightness | grep current | awk '{print $2}'`
                logger -t resume -p user.notice "Applying quirk to set full backlight brightness back to $BACKLIGHT"
                echo $BACKLIGHT > /proc/acpi/video/VID/LCD0/brightness
                ;;
        *)
                ;;
esac

[Dominik Stadler (dominik-stadler)]

My previous workaround did not fully work for some reason, however there is a suggestion at http://www.thinkwiki.org/wiki/Install_Ubuntu_8.10_%28Intrepid_Ibex%29_on_a_Thinkpad_T400#Brightness_control_broken_-_lowest_level_after_Booting.2FSuspend.2FHibernate.2FScreen-off that I will try now...

[HeadInTheClouds (jerrymweeks)]

Doesn't Connor discuss the same fix / suggestion about 7 posts above? ( "options thinkpad_acpi backlight=1")? Does this work? I have the same issue on T400.

[Stefan Hammer (j-4-deactivatedaccount)]

Hi!

I tried this suggestion and posted it on thinkwiki.
For me it works perfectly, brightness control shows same behaviour like in previous kernelversions.
With this workaround also a little bug comes back: Brightness level goes to an maximum, then jumps back and gets higher again. But this behaviour is not as annoying as the maximum dimmed screen.

Jango

[Stefan Hammer (j-4-deactivatedaccount)]

for the sake of completeness:

The workaround:

Edit following file:
sudo gedit /etc/modprobe.d/options
And add this line:
options thinkpad_acpi brightness_enable=1

Connor suggested to write "backlight=1", which does not work!

[nandhp (nandhp)]

I tried adding "options thinkpad_acpi brightness_enable=1" to /etc/modprobe.d/options and reloading the module, I also tried reloading the module with that option specified on the modprobe command, however neither approach seems to fix the problem for me. When I do "xset dpms force off" (the command to turn off the screen) the screen still returns at minimum brightness. Do I need to reboot for this to take effect?

[Stefan Hammer (j-4-deactivatedaccount)]

I did a reboot, after trying several other things, which didn't work.
After reboot, it works for me!

[Bryce Harrington (bryce)]

So far there's not been evidence to prove this is an issue with the X driver. These days X is less involved in brightness setting. The kernel is probably the more likely source of the problem. But feel free to reopen the -intel task if evidence comes to light implicating X.

[Dominik Stadler (dominik-stadler)]

I can confirm the comment from 03/04, "rmmod"ing and re-loading the module with the option did not help, only setting the option in the startup file and rebooting fixed the problem.

[Charles Profitt (cprofitt)]

This is still an issue with 2.6.27-11.31 that was updated today.

[Charles Profitt (cprofitt)]

This issue has been resolved with the 2.6.28-11-generic kernel in Jaunty.

[bigblackcar (pmioni)]

Hello.
I have the same problem on an Asus v1JP,, with ATI Mobility Radeon X1700.
I cannot move to Ubuntu 9.04 because there are graphic drivers problems.

I cannot use the
options thinkpad_acpi brightness_enable=1
fix, since mine is not a thinkpad, I think.

What should I do?