[FIXED]Last kernel update sleep/resume failure

koshak · 2013-09-19 11:31:31

UPD: This bug affects only i686 systems. Roll back and stay on linux 3.10.10 or switch to linux-lts kernel.
Follow https://bugzilla.kernel.org/show_bug.cgi?id=61781 for updates.

UPD2: Bug is fixed in linux 3.12-1.

After upgrade to linux 3.11.1-1 I get the following problem on my laptop:
when waking up from sleep I get to tty1 instead of tty7(graphical), I cannot reboot ot poweroff("Failed to find /dev/initctl" or smth like that) and in journalctl there is:

сен 19 08:50:49 mininote kernel: PM: Syncing filesystems ... done.
сен 19 08:50:49 mininote systemd[1]: Caught <SEGV>, core dump failed.
сен 19 08:50:49 mininote systemd[1]: Freezing execution.

As for systemd version, it was with 204 and with 207. After rolling back to linux 3.10.10 everything is OK.

Last edited by koshak (2013-11-22 15:54:27)

cecar · 2013-09-19 16:35:54

I have also problems resuming from suspend. I do get to desktop but the network is disabled and trying to reboot gives me the same errormessage as you. Also if i try to suspend again using systemctl suspend as root i get.

Failed to get D-Bus connection: Failed to connect to socket /run/systemd/private: Connection refused

Rolling back the kernel fixes it.

Last edited by cecar (2013-09-19 16:37:09)

venam · 2013-09-19 17:13:21

I confirm the bug.
After letting the computer sleep for sometime and waking it up I get the errors mentioned above.
Even trying to call the _old_ reboot spurts out an error.

 ~ > sudo reboot                                                                                                 
 Failed to open /dev/initctl: No such device or address
 Failed to talk to init daemon.

Last edited by venam (2013-09-19 17:24:53)

karol · 2013-09-19 17:20:03

https://bugs.freedesktop.org/show_bug.cgi?id=63189 seems similar.

Same errors as https://bbs.archlinux.org/viewtopic.php … 7#p1184467

Last edited by karol (2013-09-19 17:21:32)

fiddlinmacx · 2013-09-19 21:27:21

I'm having the same thing happen. Wow, that's annoying ;-)
I'm going to downgrade the kernel and see what happens.

agumonkey · 2013-09-20 08:57:35

Another suffering victim here.

systemd[1]: Caught <SEGV>, dumped core as pid 14587.
systemd[1]: Freezing execution.
systemd-coredump[14588]: Process 14587 (systemd) dumped core.

kernel: traps: systemd[1] general protection ip:b74b5390 sp:bfd9eadc error:0
kernel: in libc-2.18.so[b749e000+1a9000]

anecdotal digression:

I started having issues when trying lxc-{start,attach} a few days ago (can't recall if linux 3.11 was pushed). lxc force-logged me off, then systemd was all messed up. Thought it was my fault for misconfiguring lxc containers.
Recently I saw a lot of defunct/zombie processes (mainly pacman. git and zsh too, no idea why) which led me to #archlinux and #systemd where people told me to search for systemd crashes, and thanks to that here I am not alone anymore (I was starting to think my harddrive or memory was faulty)

venam · 2013-09-21 08:53:02

Did anyone test if this bug is still present with the systemd version that just came out?

falconindy · 2013-09-21 10:06:25

venam wrote:

Did anyone test if this bug is still present with the systemd version that just came out?

This isn't a systemd bug... It's a kernel bug which is corrupting memory and causing systemd to segfault.

agumonkey · 2013-09-21 12:02:58

venam wrote:

Did anyone test if this bug is still present with the systemd version that just came out?

$ pacman -Qi systemd
Name : systemd
Version : 207-5

With 207-5 I did a suspend/resume and I don't have the problem anymore.

falconindy wrote:

venam wrote:
Did anyone test if this bug is still present with the systemd version that just came out?
This isn't a systemd bug... It's a kernel bug which is corrupting memory and causing systemd to segfault.

Do you have a ml or bug report link saying this ? I couldn't find any explanation, and as I said to @venam, for now, systemd 207-5 seems to "make a difference".

wkatastrof · 2013-09-21 14:00:10

Greetings,

I had this problem up until yesterday when I downgraded kernels. I still had to hard shutdown after the downgrade 'cause I didn't know how to recover systemd. Now it seems to be ok.

If there's any info I can post for you guys, I'd be glad too, just let me know. I'm still learning about systemd and how it can let me monitor my system.
I saw they moved from GPL to LGPL which kind of sucks for free software, but there's nothing to do about it I guess...

Anyway, let me know if I can be of assistance.

georgej116 · 2013-09-22 23:55:16

I can confirm also. Since updating yesterday, whenever my laptop (Lenovo Ideapad s10e) wakes up from suspend, systemd is gone and I can't even shutdown properly.

darius · 2013-09-23 06:38:47

Same problem here. After a suspend, my network interfaces are unmanaged and I can't shutdown properly.
I am running systemd 207-5. The problem still persists.

kmiernik · 2013-09-23 06:41:24

Yet another confirmation. My computer is Asus Eee 1005P. With linux-3.11.1-1-i686 sleep/resume causes the systemd to be dead. The following error message is found in the kernel log:

kernel: [ 1962.265804] traps: systemd[1] general protection ip:b74dcee3 sp:bfe79fe8 error:0 in libc-2.18.so[b745f000+1a9000]

In the linux-3.10.9-1-i686 kernel this bug is not present.

emulti · 2013-09-23 08:34:42

Confirmed here with linux-pae-3.11.1-1 and systemd 207-5

The problem does not occur if the suspend time is short. It only appears after system sleep for an hour or two, or overnight.

On wakeup systemd crashes, and the connection to dbus fails:
Sep 23 08:48:28 emulti1 systemd[1183]: Time has been changed
Sep 23 08:48:28 emulti1 kernel: [ 342.038653] traps: systemd[1] general protection ip:b7546070 sp:bf9b7000 error:0 in libc-2.18.so[b74d5000+1a9000]
Sep 23 08:48:28 emulti1 systemd-sleep[3088]: System resumed.
Sep 23 08:48:28 emulti1 systemd[1]: Caught <SEGV>, dumped core as pid 3170.
Sep 23 08:48:28 emulti1 systemd[1]: Freezing execution.
Sep 23 08:48:28 emulti1 systemd-coredump[3176]: Process 3170 (systemd) dumped core.
...
Sep 23 08:49:20 emulti1 dbus[699]: [system] Failed to activate service 'org.freedesktop.systemd1': timed out
Sep 23 08:49:20 emulti1 dbus-daemon[699]: dbus[699]: [system] Failed to activate service 'org.freedesktop.systemd1': timed out

I'm going to try rolling back to previous kernel version.

kmiernik · 2013-09-23 08:46:45

emulti wrote:

The problem does not occur if the suspend time is short. It only appears after system sleep for an hour or two, or overnight.

In my case I haven't observed correlation with suspension time. My systemd dies even if I suspended system for 1.9 s (hit the suspend button, wait until fully suspended, hit enter to wake up) as well as half an hour.

fethio · 2013-09-23 18:24:14

same here. I've got a Lenovo 3000 N100 and upon

 ~ > sudo reboot                                                                                                 
 Failed to open /dev/initctl: No such device or address
 Failed to talk to init daemon.

Strike0 · 2013-09-23 22:12:27

It seems that only the i686 kernel is affected (anyone with 64 bit experiencing it in this thread?).

Relevant kernel bug to add info to: https://bugzilla.kernel.org/show_bug.cgi?id=61781

cecar · 2013-09-23 22:26:12

My machine with 64bit kernel works perfectly but my 32bit machine does not so i guess its only the 32bit kernel that has this bug.

georgej116 · 2013-09-24 12:14:57

Same here. Problem only occurs on my 32-bit machine, and not the 64-bit. Both are running kernel 3.11.1-1 and systemd 207-5.

I can also confirm that reverting to the 3.10 kernel fixes the problem.

Last edited by georgej116 (2013-09-24 12:26:02)

awagner · 2013-09-24 12:25:27

and here: https://bbs.archlinux.org/viewtopic.php?id=170111

thesofty · 2013-09-24 17:34:29

I also have the suspicion that only the 32bit version has the problem. Currently, I've install two Lenovos with Arch: An old X61 with 32bit that one has the same problems after resume, and a X1 with 64bit that works perfectly with Kernel 3.11.1 and Systemd 205.
I also can confirm that downgrading the 32bit-machine to kernel 3.10.10 let disappear the issue.

Update:
Meanwhile, I currently have installed Systemd 207 on both systems.

Last edited by thesofty (2013-09-24 17:35:55)

abilng · 2013-09-25 07:57:42

Same problem here.
I am getting following error when I try to do any thing with systemd,after my system resumes from sleep.:

Failed to get D-Bus connection: Failed to connect to socket /run/systemd/private: Connection refused

agumonkey · 2013-09-25 09:49:08

I forgot to mention I'm running a 32bit linux 3.11.1-1 too. But I had no visible issue since systemd 207-5. Going from normal state to suspend is uber fast (~ second total), even night-long sleeps doesn't cause bugs. The 3.11.1-2 kernel has just been pushed, I'll see if that changes things, for better or worse.

update : no changes with 3.11.1-2

Last edited by agumonkey (2013-09-25 10:42:01)

cecar · 2013-09-25 09:53:16

The new kernel is not fixing this issue for me at least.

WorMzy · 2013-09-25 11:46:59

The -2 linux package just has a fix for a skge (network) module regression. Like Aaron Lu says in the bug report, he can't reproduce the bug, and needs some extra information to diagnose the problem. Jeff Pohlmeyer looks to have done most of the hard work and isolated a cluster of commits where the problem surfaced. If you can find the time to assist with the investigation and share your findings on the bug report, it may help find a fix sooner.

Last edited by WorMzy (2013-09-25 11:47:17)

Arch Linux

#1 2013-09-19 11:31:31

[FIXED]Last kernel update sleep/resume failure

#2 2013-09-19 16:35:54

Re: [FIXED]Last kernel update sleep/resume failure

#3 2013-09-19 17:13:21

Re: [FIXED]Last kernel update sleep/resume failure

#4 2013-09-19 17:20:03

Re: [FIXED]Last kernel update sleep/resume failure

#5 2013-09-19 21:27:21

Re: [FIXED]Last kernel update sleep/resume failure

#6 2013-09-20 08:57:35

Re: [FIXED]Last kernel update sleep/resume failure

#7 2013-09-21 08:53:02

Re: [FIXED]Last kernel update sleep/resume failure

#8 2013-09-21 10:06:25

Re: [FIXED]Last kernel update sleep/resume failure

#9 2013-09-21 12:02:58

Re: [FIXED]Last kernel update sleep/resume failure

#10 2013-09-21 14:00:10

Re: [FIXED]Last kernel update sleep/resume failure

#11 2013-09-22 23:55:16

Re: [FIXED]Last kernel update sleep/resume failure

#12 2013-09-23 06:38:47

Re: [FIXED]Last kernel update sleep/resume failure

#13 2013-09-23 06:41:24

Re: [FIXED]Last kernel update sleep/resume failure

#14 2013-09-23 08:34:42

Re: [FIXED]Last kernel update sleep/resume failure

#15 2013-09-23 08:46:45

Re: [FIXED]Last kernel update sleep/resume failure

#16 2013-09-23 18:24:14

Re: [FIXED]Last kernel update sleep/resume failure

#17 2013-09-23 22:12:27

Re: [FIXED]Last kernel update sleep/resume failure

#18 2013-09-23 22:26:12

Re: [FIXED]Last kernel update sleep/resume failure

#19 2013-09-24 12:14:57

Re: [FIXED]Last kernel update sleep/resume failure

#20 2013-09-24 12:25:27

Re: [FIXED]Last kernel update sleep/resume failure

#21 2013-09-24 17:34:29

Re: [FIXED]Last kernel update sleep/resume failure

#22 2013-09-25 07:57:42

Re: [FIXED]Last kernel update sleep/resume failure

#23 2013-09-25 09:49:08

Re: [FIXED]Last kernel update sleep/resume failure

#24 2013-09-25 09:53:16

Re: [FIXED]Last kernel update sleep/resume failure

#25 2013-09-25 11:46:59

Re: [FIXED]Last kernel update sleep/resume failure

Board footer