Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

ulukai · 2009-12-30 16:30:54

Today I updated my laptop with the new 2.6.32.2-2 kernel from core. Previous update of my system was just a few days back, so there are no massive changes to other packages and I also don't use testing packages.

Since the update, after closing the laptop lid (which just blanks my screen) and I want to continue working I just have a black screen (no backlight) and the screen won't respond to mouse or keyboard input. I have to reboot (ctr-alt-del or pushing my power button) or switch to VT1 and back to VT7 to get my screen back. I'm absolutely positive that this was not a problem before, it worked fine with 2.6.30 and 2.6.31 kernels with my same configuration.

HP 6730b laptop
Intel GMA 4500 (Corporation Mobile 4 Series Chipset Integrated Graphics Controller (rev 07))
Intel driver 2.9.1-1 using KMS with early start (modprobe.conf) and I rebuilt my initcpio to the new 2.6.32.2-2 kernel
Gnome 2.28.2

Can anyone confirm this behaviour? Should I file a bug somewhere? I read about lots of upgrade problems with this new kernel, but for the rest everything else works fine so far.

Thanks for your feedback.

Ulukai

martusia · 2009-12-30 16:38:07

I confirm it with my Dell D420. Exactly the same problem, though, if I have tty1 active I can type reboot and it responds, so looks like system is working though with screen switched off.
But still, I confirm.

ulukai · 2009-12-30 16:52:37

martusia wrote:

I confirm it with my Dell D420. Exactly the same problem, though, if I have tty1 active I can type reboot and it responds, so looks like system is working though with screen switched off.
But still, I confirm.

Indeed, system still responds, but the screen stays black (no backlight) until I switch VT and switch back OR reboot. Do you also use KMS?

Would it be interesting if I ask you and others who experience the same problem to say which power controller they user? I have Gnome-Power-Manager, but maybe it is not a problem in KDE's PowerDevil backend? Or would it really be a kernel / KMS issue?

Also, I noticed that there has been work done at this subject, but for me it really is a regression:
http://git.kernel.org/?p=linux/kernel/g … c39192d6cc

Last edited by ulukai (2009-12-30 16:53:51)

kheophex · 2009-12-30 17:06:02

Hello,

I confirm the exact same problem with my DELL Latitude E5400, which comes with an intel card. Otherwise the new kernel works just fine with KMS. I'm using laptop-mode tools as the powermanager and a script which suspends to ram when the lid is closed. However when the lid is lifted again the laptop display is not turned back on even though everything else seems to be working.

I think that's it's somehow connected to xrandr though. If I close the laptop lid when the laptop is connected to another LCD monitor, when resuming operation the LCD monitor springs back to life but not the main display. If I switch from to ttys then I see both displays lit up in that strange clone mode. The main display can be turned back on, atleast on my system, with a

 xrandr --output LVDS1 --auto --right-of VGA1

(your displaynames and positions will vary) . Interestingly enough, even if I disable the LCD monitor with xrandr and then close and lift the lid what happens is that the LCD monitor will be the main monitor whilst the main laptop display is turned off (and the desktop does not extend beyond this monitor).

I hope this might give some extra clue.

moose jaw · 2009-12-31 02:47:39

Have you tried any of the pm-suspend quirks? On my Samsung NC10 (Atom netbook, Intel 945GME, with Intel 2.9.1 driver and early-start KMS as described on the wiki), the 2.6.32 update initially killed my screen after resuming from suspend-to-RAM. No response from the keyboard, no hardward light from caps lock, etc: only a hard reboot would work. Adding the "--quirk-vbestate-restore" flag to pm-suspend has fixed the problem for me.

ulukai · 2009-12-31 08:30:40

@ moose jaw: What is this pm-suspend quirk? How can I test if this makes any difference and does it only matter when resuming after suspending, or also when closing the lid just turns off the backlight? When my lid is opened again, my hardware light for caps lock DOES work.

@ kheophex: I don't use multiple monitors but will see if I can find an opportunity to test this.

Also I noticed that my monitor isn't being turned off after 10 minutes of idle time. This worked before the update, as configured in gnome-power-manager...

moose jaw · 2009-12-31 15:38:56

ulukai wrote:

@ moose jaw: What is this pm-suspend quirk? How can I test if this makes any difference and does it only matter when resuming after suspending, or also when closing the lid just turns off the backlight? When my lid is opened again, my hardware light for caps lock DOES work.

You can read about all the quirks in the pm-suspend man page ("man pm-suspend"). I don't use Gnome, so I don't know what program the Gnome power manager calls (or what acpid configuration it relies on, or whatever) to suspend the machine. I'm running Openbox and have set a key combo to run pm-suspend, but you can just as easily test it out by running it in a terminal. You need to be root (or set yourself up for sudo; check the wiki page on pm-utils) to run pm-suspend. To use a quirk, instead of entering "sudo pm-suspend" at the prompt, try, e.g., "sudo pm-suspend --quirk-vbestate-restore". There is no command to enter upon resuming, but the different quirks cause pm-suspend to save different types of information when suspending, and then that special info/state/whatever can be used by the system when resuming, which in some cases can be helpful.

youplease · 2009-12-31 16:28:27

I have the same problem on my Dell Inspiron 1525 (intel graphics, early KMS start). After upgrading to 2.6.32 when I close the lid and then open it the screen remains black.

Last edited by youplease (2009-12-31 18:12:30)

ulukai · 2010-01-01 22:43:44

Happy New Year everyone :-)

I've got some more news. Tested it also on my other laptop, which is an IBM Thinkpad R50e. It has an older Intel Graphics chip (855 I guess) and I also have KMS enabled on this system. Today I tested it with 2.6.31 + KMS, which works fine. Then I upgraded to 2.6.32 + KMS and it still worked: closing the screen turns off the backlight and opening the lid turns it on again. The only differences are that on that system I use KDE 4.3.4 instead of Gnome and it is an older Intel chip. The KMS setup and driver however, is exactly the same.

Apparently GPM (gnome-power-manager) depends on devicekit-power, which on its turn depends on pm-utils. KDE however uses PowerDevil, it's own backend which is a part of KDE.

Seems to me that it is either related to the backend, the support of the intel driver for various chipsets or a KMS problem. Can someone who also reported this bug try disabling KMS and see of this is still a problem? If not, I will try it myself next week, I have a busy chedule right now... ;-)

@ Moose Jaw: I also looked into pm-suspend, but i guess it is only usefull for suspending and resuming sessions (something I never do)? Is there a command you know for simulating closing the lid and reopening it again?

In the meantime, should we open a bug already with Intel or Gnome?

mezcal · 2010-01-02 09:08:28

Hi,

I have the same problem. Laptop HP nx6110 with intel chipset 910GM.
I use XFCE4 and don`t use KMS.

I tried disable suspending after closing lid. The screen stay black after opening it.

Kernel 2.6.31 works fine.
I think it`s kernel or driver bug.

Habmala · 2010-01-02 09:35:21

Hi

I have the same issue, after closing the laptops lid it shuts of the screen as it's supposed to. But when opening the lid it flashes and goes blank, it's not turnd of, but black.

I't worked fine with 2.6.31 and no KMS, now on 2.6.32 and the default KMS turned on it fails.

I also have an intel card, mine is in a "toshiba satellite L40 -18R" and I run fluxbox in X so I don't think it's a GNOME issue.

When I close and open the lid and go to tty1 (ctrl+alt+F1) it sais "Warning output VGA not found; ignoring" I will try to se later if this is still there after pluging in an external monitor to the VGA output, I don't know if it used to be there before, but I don't think so.

Stalafin · 2010-01-02 11:52:57

Looks as if a solution was suggested in this thread:
http://bbs.archlinux.org/viewtopic.php?id=87828

Wen · 2010-01-03 02:10:25

Happy new year everyone!

I don't think this is a problem of X11 or any WM. I have an Arch Linux box with only console, which, after being upgraded, has the same problem. The box has an Intel 855GM integrated graphic card. May we assume this is a bug (of kernel?) with regard to Intel graphic chips?

moose jaw · 2010-01-03 02:46:14

ulukai wrote:

@ Moose Jaw: I also looked into pm-suspend, but i guess it is only usefull for suspending and resuming sessions (something I never do)? Is there a command you know for simulating closing the lid and reopening it again?

Isn't suspending and resuming a session exactly what you're doing (or hoping to do) when you close your lid and then open it up again? I.e., putting the machine into a low-power sleep state and then having everything come back to where you were when you wake it up? You mentioned that the Gnome power management tools depend on pm-utils, which suggests that somewhere in there "pm-suspend" is getting called. If you can find out where it's getting called (i.e., in some script that Gnome power manager uses), and try changing it to "pm-suspend --quirk-vbestate-restore" (or one of the other quirks, perhaps), then that may bring your screen back on wakeup.

ulukai · 2010-01-03 11:17:36

moose jaw wrote:

ulukai wrote:
@ Moose Jaw: I also looked into pm-suspend, but i guess it is only usefull for suspending and resuming sessions (something I never do)? Is there a command you know for simulating closing the lid and reopening it again?
Isn't suspending and resuming a session exactly what you're doing (or hoping to do) when you close your lid and then open it up again? I.e., putting the machine into a low-power sleep state and then having everything come back to where you were when you wake it up? You mentioned that the Gnome power management tools depend on pm-utils, which suggests that somewhere in there "pm-suspend" is getting called. If you can find out where it's getting called (i.e., in some script that Gnome power manager uses), and try changing it to "pm-suspend --quirk-vbestate-restore" (or one of the other quirks, perhaps), then that may bring your screen back on wakeup.

Hehe, nope I don't hope to suspend :-) I just close the lid to save the display (and some power) but want my machine to keep running (when downloading torrents or when I'm going to eat etc...) Also it keeps the dust from entering the keyboard ;-)

I also took a look at the post where there should be a patch present, but I think it's for another problem: flickering screens.

For me it's a bug in intel driver or kernel, but still it's strange that KDE doesn't suffer from this!

moose jaw · 2010-01-03 15:34:43

ulukai wrote:

Hehe, nope I don't hope to suspend :-) I just close the lid to save the display (and some power) but want my machine to keep running (when downloading torrents or when I'm going to eat etc...) Also it keeps the dust from entering the keyboard ;-)

Aha, my mistake. Well, if you decide you want to start suspending with 2.6.32, now you have some tips! Good luck with the screen backlight problem.

brianhanna · 2010-01-03 15:48:04

I have Intel 945GM graphics and I'm not having this problem with KDE. Haven't tried any other WMs.

jlacroix · 2010-01-03 15:57:29

Hello everyone. I have a Dell Inspiron 15n (the one that shipped with Ubuntu) and I think I might be having this problem. I set KDE to go into suspend when closing the lid. When I do that, it never suspends. When I open the lid to check on it, I have a plain black screen with a white cursor on the top left. I cannot get the screen back unless I hold down the power button and forcefully cut it off. Is this the same issue as the one in this topic?

kheophex · 2010-01-03 22:41:24

Hello,

Looking at http://patchwork.kernel.org/patch/62734/ it seems that it's a kernel 2.6.32 issue that you can potentially solve by patching your kernel sources and recompiling. It's not much of deal to me anyway, so I've just set a shortcut in my XFCE config to issue a

xrandr --output LVDS1 --auto

when I press a key combination (Super + 1 for me) to turn the monitor back on (xrandr works even if you're not using dual head). I suppose that patch will probably get into the mainline soon.

Kind Regards.

ulukai · 2010-01-04 10:03:37

kheophex wrote:

Hello,
Looking at http://patchwork.kernel.org/patch/62734/ it seems that it's a kernel 2.6.32 issue that you can potentially solve by patching your kernel sources and recompiling. It's not much of deal to me anyway, so I've just set a shortcut in my XFCE config to issue a
xrandr --output LVDS1 --auto
when I press a key combination (Super + 1 for me) to turn the monitor back on (xrandr works even if you're not using dual head). I suppose that patch will probably get into the mainline soon.
Kind Regards.

Thanks for your post, now we know it is kernel related and there is a patch provided. Hopefully it gets in the mainline and arch kernel soon!
I'm not going to recompile my kernel for this little problem, but it's nice to know for those who want to.

jlacroix · 2010-01-04 13:30:12

ulukai wrote:

kheophex wrote:
Hello,
Looking at http://patchwork.kernel.org/patch/62734/ it seems that it's a kernel 2.6.32 issue that you can potentially solve by patching your kernel sources and recompiling. It's not much of deal to me anyway, so I've just set a shortcut in my XFCE config to issue a
xrandr --output LVDS1 --auto
when I press a key combination (Super + 1 for me) to turn the monitor back on (xrandr works even if you're not using dual head). I suppose that patch will probably get into the mainline soon.
Kind Regards.
Thanks for your post, now we know it is kernel related and there is a patch provided. Hopefully it gets in the mainline and arch kernel soon!
I'm not going to recompile my kernel for this little problem, but it's nice to know for those who want to.

Thank you. I hope that it does make it into the Arch kernel as well. I am new to Arch so I'm not sure if that's something the Arch developers typically do, but if they did it would be awesome.

ulukai · 2010-01-07 12:59:57

Actually I don't think Arch devs patch the kernel. They use upstream sources so my guess is that upstream has to build it in the kernel. Hopefully they don't skip it for the 2.6.32 series as I think I've read it will be implemented in 2.6.33...

jlacroix · 2010-01-07 13:21:37

ulukai wrote:

Actually I don't think Arch devs patch the kernel. They use upstream sources so my guess is that upstream has to build it in the kernel. Hopefully they don't skip it for the 2.6.32 series as I think I've read it will be implemented in 2.6.33...

I hope upstream patches it into the kernel then. That would make me very happy. I'm going to leave it alone and wait for an update, but if an update doesn't happen I guess I will have to patch it.

jithin1987 · 2010-01-07 16:32:03

In my case with kde 4.3.4 and kernel 2.6.32 I am seeing this sometimes. But most times it just works. I have noticed this 3 times so far.

I saw these in my errors.log

Jan  7 21:11:16 grainlead-lx kernel: BUG: soft lockup - CPU#0 stuck for 61s! [X:5277]
Jan  7 21:12:21 grainlead-lx kernel: BUG: soft lockup - CPU#0 stuck for 61s! [X:5277]
Jan  7 21:13:27 grainlead-lx kernel: BUG: soft lockup - CPU#0 stuck for 61s! [X:5277]
Jan  7 21:13:56 grainlead-lx kernel: INFO: task i915/1:22223 blocked for more than 120 seconds.
Jan  7 21:13:56 grainlead-lx kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
Jan  7 21:14:32 grainlead-lx kernel: BUG: soft lockup - CPU#0 stuck for 61s! [X:5277]
Jan  7 21:15:38 grainlead-lx kernel: BUG: soft lockup - CPU#0 stuck for 61s! [X:5277]
Jan  7 21:15:56 grainlead-lx kernel: INFO: task i915/1:22223 blocked for more than 120 seconds.
Jan  7 21:15:56 grainlead-lx kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
Jan  7 21:16:43 grainlead-lx kernel: BUG: soft lockup - CPU#0 stuck for 61s! [X:5277]

I think I have lost track the last time I used a fully functional kernel. Starting from 28 I have been anticipating a new version of kernel because of one fix or other

Last edited by jithin1987 (2010-01-07 16:38:38)

Drakebyte · 2010-01-08 14:24:39

DELL laptop owners should try this:

open /boot/grub/menu.lst with your editor of choice, as root, and find:

# (0) Arch Linux
title Arch Linux
root (hd0,0)
kernel /vmlinuz26 root=/dev/disk/by-uuid/e9e3ecc2-2c13-4251-9151-84ef7971d78c ro noapic
initrd /kernel26.img

Just add the part in bold, that fixed it for me at least. The non-bold part changes from installation to installation, but it's the first entry after all the comments.

Good luck,
D

Last edited by Drakebyte (2010-01-08 14:31:13)

Arch Linux

#1 2009-12-30 16:30:54

Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#2 2009-12-30 16:38:07

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#3 2009-12-30 16:52:37

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#4 2009-12-30 17:06:02

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#5 2009-12-31 02:47:39

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#6 2009-12-31 08:30:40

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#7 2009-12-31 15:38:56

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#8 2009-12-31 16:28:27

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#9 2010-01-01 22:43:44

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#10 2010-01-02 09:08:28

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#11 2010-01-02 09:35:21

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#12 2010-01-02 11:52:57

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#13 2010-01-03 02:10:25

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#14 2010-01-03 02:46:14

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#15 2010-01-03 11:17:36

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#16 2010-01-03 15:34:43

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#17 2010-01-03 15:48:04

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#18 2010-01-03 15:57:29

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#19 2010-01-03 22:41:24

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#20 2010-01-04 10:03:37

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#21 2010-01-04 13:30:12

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#22 2010-01-07 12:59:57

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#23 2010-01-07 13:21:37

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#24 2010-01-07 16:32:03

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

#25 2010-01-08 14:24:39

Re: Kernel 2.6.32.2-2 black screen after closing laptop lid (intel + KMS)

Board footer