[Lenovo L420] System unable to resume from suspend

[Marc Legris]

Binary package hint: pm-utils

System is unable to return to the desktop from suspend mode. Resuming will instead show only a blank screen.

ProblemType: Bug
DistroRelease: Ubuntu 11.04
Package: pm-utils 1.4.1-7
ProcVersionSignature: Ubuntu 2.6.38-7.39-generic 2.6.38
Uname: Linux 2.6.38-7-generic i686
Architecture: i386
Date: Fri Apr 8 10:49:13 2011
InstallationMedia: Ubuntu 11.04 "Natty Narwhal" - Beta i386 (20110330)
ProcEnviron:
 LANGUAGE=en_US:en
 LANG=en_US.UTF-8
 SHELL=/bin/bash
SourcePackage: pm-utils
UpgradeStatus: No upgrade log present (probably fresh install)

[Chris Van Hoof (vanhoof)]

Please confirm on the 2.6.38-8 kernel available in -updates

[Martin Pitt (pitti)]

This system uses an Intel graphics chip and thus uses KMS, so pm-utils does not and should not apply any quirks. Moving to kernel.

[Victor Tuson Palau (vtuson)]

Marc - can you retest with beta2 plz

[Marc Legris (maaarc-deactivatedaccount-deactivatedaccount-deactivatedaccount-deactivatedaccount-deactivatedaccount)]

Updated with the latest packages, issue is still present.

[Ara Pulido (ara)]

Assigning back to Vanhoof

[Seth Forshee (sforshee)]

@Marc:

Have you done any tests to see whether the machine is hung when this happens or the screen just isn't coming on? For instance, have you tried pinging the machine on the network when it is in this state?

If the machine is still alive, I'd suggest using ssh to log into the machine while it is in this state and running 'apport-collect 754712' to attach relevant logs to this bug report. If it is unresponsive, try some of the steps in the following links to see if you can get any information about where things are going bad.

https://wiki.ubuntu.com/DebuggingKernelSuspend
https://wiki.ubuntu.com/DebuggingKernelSuspendHibernateResume

Thanks!

[Jeremy Foshee (jeremyfoshee)]

milestoned for Natty updates.

~JFo

[actionparsnip (andrew-woodhead666)]

Also make sure you have latest BIOS

[Victor Tuson Palau (vtuson)]

assigning to Manjo as vanhoof is away

[dddolce (dolce-donatello)]

Same problem with Lenovo T410, expecially when I have a secondary display connected.

[David Colebatch (david-subdata)]

Same problem with ASUS UL30A

I can see and move the cursor, but the screen is blank. My laptop's brightness keys, as well as capslock all work. I cannot ctrl+alt+f1 to a console though.

[drew (latortuga)]

I am also experiencing this issue on an ASUS laptop, the U43JC. When I suspend, the screen goes blank and the machine becomes unresponsive to anything except a long-press on my power button.

[Chris Van Hoof (vanhoof)]

@marc -- Can you verify with the latest natty -proposed kernel? There have been a few fixes in this area, and given that this is using standard sandybridge UMA graphics, it should suspend properly.

Cheers,
Chris

[Marc Legris (maaarc-deactivatedaccount-deactivatedaccount-deactivatedaccount-deactivatedaccount-deactivatedaccount)]

@Seth

System does not respond to ssh when resuming from suspend. I did see the magic number and hash matches in desg.

Keep in mind this system is also affected by: Bug #777212

[Marc Legris (maaarc-deactivatedaccount-deactivatedaccount-deactivatedaccount-deactivatedaccount-deactivatedaccount)]

@Chris

Updating to 2.6.38-9 does not fix the resume issue.

[pinguaw (aweiser)]

Hello Guys,

I have the same problem with my Lenovo U160 Ideapad. Tried to unload all video modules before suspend. Added a file in /etc/pm/config.d with the following content:

SUSPEND_MODULES="video i2c_algo_bit drm drm_kms_helper i915 videodev uvcvideo"

But no success, same after resume screen keeps blank but illuminated. The system is responsive ( I can ping it) and even another suspend (Fn + F1) works very well. The pm-suspend.log does not state any error. But I also do not see to unload any module (does the variable SUSPEND_MODULES do the job??).

I really need this suspend feature, how can I help? I have good developer knowledge, but need some guidance to get the right information where to look (take to long to find this out for my self). Let me know what I should try or debug;-)

I have Ubuntu 11.04 with kernel 2.6.38-8-generic-pae i686 (32bit) fresh new installation.

[HwRhymes (hanswurst-deactivatedaccount)]

This might give hints on what to do about it:
https://bugs.freedesktop.org/show_bug.cgi?id=35462

Supposedly, it is related to a faulty BIOS.

[Seth Forshee (sforshee)]

@HwRhymes: The bios certainly could be to blame, but resume hangs are extremely hardware specific, so unless the commentor on the freedesktop bug you linked to had an identical machine it's hard to say whether or not it's the same problem.

@Marc: Something is getting stored in the resume trace but the kernel could't find the matching device. One possibility is that the hang was in a kernel module that wasn't loaded when the pm_trace output was generated during boot. Could you try the pm_trace again, but this time after fully booted run 'cat /sys/power/pm_trace_dev_match' and paste the output here? Thanks!

[Mariano Loza (mariano-loza)]

I have the same bug after a fresh install in a Lenovo v360.
when resuming after suspend, I can even hear the music of a video I was watching, but I am not able to see anything.

on dmseg I find this:

2.322986] i2c-core: driver [adp5520] using legacy suspend method
[ 2.322988] i2c-core: driver [adp5520] using legacy resume method

Mariano

[Seth Forshee (sforshee)]

@Mariano: Yours sounds like a different bug then, as Marc reported that his system is completely unresponsive after trying to resume. In your case it sounds like the system resumes but something goes wrong with the graphics. Please run 'ubuntu-bug linux' in a terminal to open a new bug for your issue.

[pinguaw (aweiser)]

Hi,

cat /sys/power/pm_trace_dev_match

gives only:

block

This is all. I built my own kernel with acpi debuging on. I will attach dmesg maybe you will see an interesting line.

[Helge Willum Thingvad (helgesdk)]

I am sometimes having a similar symptom yet possibly different bug on my ThinkPad X200 with Intel 4500GMA graphics.

Once in a while, when the system resumes from suspend, the screen will be blank with only the cursor showing. It seems that the desktop is still somewhat responsive; even though I can't see anything, the cursor changes shape as I move it around and I can type commands.

Just now I experienced this issue, but I was able to Ctrl+Alt+F1 to the console. The desktop would remain blank with just the cursor showing if I switched back to TTY8.
.
I managed to restore the desktop by killing compiz (compiz --replace wasn't enough) then re-running it on DISPLAY :0.
$ pkill -9 compiz
$ DISPLAY=:0 compiz

I am using the Ubuntu Classic session.

Should I post a new bug report? If so, should I do it via apport-bug?

[Seth Forshee (sforshee)]

@Helge, that's definitely a different problem. Please open a new bug if you cannot find any existing bug matching your symptoms. apport-bug should be fine for opening the new bug.

[Chris Van Hoof (vanhoof)]

Hi Marc -- Please see Seth's updates in Comment #21

Regards,
Chris

[Marc Legris (maaarc-deactivatedaccount-deactivatedaccount-deactivatedaccount-deactivatedaccount-deactivatedaccount)]

@Chris --output shows the following:
block

pnp

[Seth Forshee (sforshee)]

@Marc, thanks. Can you also provide the output of lsmod from this machine?

[Seth Forshee (sforshee)]

Does removing all the tpm modules before suspending help?

[Marc Legris (maaarc-deactivatedaccount-deactivatedaccount-deactivatedaccount-deactivatedaccount-deactivatedaccount)]

@Seth -- removing all the tpm modules does not allow the system to resume normally.

[Chris Van Hoof (vanhoof)]

Marc -- What BIOS revision is your L420 presently on, I'm curious to see what changes may have been made in later releases

[Marc Legris (maaarc-deactivatedaccount-deactivatedaccount-deactivatedaccount-deactivatedaccount-deactivatedaccount)]

Chris -- Version is 8GET26WW (1.05)

[Chris Van Hoof (vanhoof)]

Marc -- Might be worth investigating an update to 1.07 as it seems like the most substantial update so far:

http://download.lenovo.com/ibmdl/pub/pc/pccbbs/mobiles/8guj03uc.txt

[Marc Legris (maaarc-deactivatedaccount-deactivatedaccount-deactivatedaccount-deactivatedaccount-deactivatedaccount)]

Chris -- Updating to 1.07 does not fix any of the system errors for the L420

[Colin Ian King (colin-king)]

@Marc, debugging this is not going to be easy. Can I get access to this machine somehow?

[Marc Legris (maaarc-deactivatedaccount-deactivatedaccount-deactivatedaccount-deactivatedaccount-deactivatedaccount)]

Using the steps from here: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/777212/ allows the system to suspend normally.

[Chris Van Hoof (vanhoof)]

Assigning Timo to see why forcing KMS seems to sort things out here on the L420

[Timo Aaltonen (tjaalton)]

From bug 777212:

"what if you add 'GRUB_GFXPAYLOAD_LINUX=text' to /etc/default/grub (run update-grub to update the actual config) and reboot, does it change anything?"

"Could you also try the latest mainline kernels from http://kernel.ubuntu.com/~kernel-ppa/mainline/" (it was reported that 2.6.39-r4 fixed that bug)

[Marc Legris (maaarc-deactivatedaccount-deactivatedaccount-deactivatedaccount-deactivatedaccount-deactivatedaccount)]

Timo -- both those options allows the system to suspend