Page MenuHomeSolus

Spontaneous Log Offs
Open, Needs More InfoPublic

Description

I have for at least two weekly update cycles been experiencing an issue where I will suddenly just be logged off while using my laptop, which closes all my applications and is generally quite frustrating. I suppose I don't for sure know that it is a log off that is happening, the screen will suddenly go black for a couple seconds, and then I will be presented with the login screen (not the lock screen), and upon logging in all applications will have been closed which does mirror the exact behavior when I log off manually. I have fully updated my system every week when the new package updates get pushed. I would estimate that it has happened in total about 5 times in the last two or maybe three weeks, sorry that I am not able to remember exactly when it started happening. The last two times it happened was within 10 minutes of waking from suspend, not sure if that is significant, I can't see any other pattern in when it is happening based on what applications are running or anything. I can provide whatever logs may be helpful to see, but I am not familiar enough with Solus to know exactly what would be necessary to see. Also unclear if this is the correct place to post this and am happy to copy the discussion elsewhere.

System Information:
Dell G5 5590 Laptop
i7-9750H CPU
GeForce GTX 1650 Mobile / Max-Q using the latest NVIDIA drivers, including the 32 bit drivers
uname -a output: 5.10.15-172.current #1 SMP PREEMPT Fri Feb 12 16:35:57 UTC 2021 x86_64 GNU/Linux

Event Timeline

Hi @jm-brennan , I think I have the same issue since an update in early Janunary 2021.
Had some discussion about this problem in the Solus forum https://discuss.getsol.us/d/2616-laptop-doesn-t-properly-turn-on-after-suspend/, but no solution found.
It still seems to me that some applications like VS Code or Brave browser increase the probability that this happens, but without it also sometimes logs me out spontaneously shortly after wakeup from suspend.

System information:
Sony Vaio VPC-EB3X
Intel CoreTM i5-460M
ATI MobilityRadeon HD5650

All latest updates installed, problem occurs with both current and LTS kernels.

palto42 added a comment.EditedMar 21 2021, 1:13 PM

I have found a similar issue in Arch Linux forum, which seems to be solved by installing bbswitch which is available in Solus reop.
https://bbs.archlinux.org/viewtopic.php?id=244635

Tried it, but doesn't fix this problem.

Sounds like Xorg might be crashing on resume and dumping you back to the DM to log in again. Logs?

palto42 added a comment.EditedMar 21 2021, 6:42 PM

Some logs from one of todays' forced logoff's:

https://dev.getsol.us/F7219923

I ran journalctl with -p 7 but still am not seeing the same level of detail that @palto42 's log has, I'm not familiar with journalctl so if there is a way to get more detailed logs just let me know how and I will try it, otherwise here is a log from one that happened today:

https://dev.getsol.us/F7222909

palto42 added a comment.EditedMar 22 2021, 5:53 PM

Hi @jm-brennan , I'm simply running sudo journalctl -S -2h to see all detail. The start time is defined with the option -S -2h (since 2h ago), you can also specify an absolute date like -S "2021-03-16 20:00:00".

You can also double-check the systemd log-level with systemctl log-level which is default info on my PC.

szb added a subscriber: szb.Apr 17 2021, 5:30 PM

This behaviour affects me to as of this week. I've attached the output of journalctl for the last hour before the last involuntary logout occured. Since I only have integrated Intel graphics, I won't bother to install bbswitch as @palto42 tried. If I can provide better/more detailed information, please let me know how.

The affected machine is an
Acer Aspire A5,
Intel i5-8250U CPU
Intel UHD Graphics 620

Thanks for your time!

szb added a comment.Apr 19 2021, 5:37 PM

I was just again logged out of the session, this time the machine basically just sat idly by, with only one tab of Firefox running. For the purpose of (hopefully) tracking down the error, I will continue to upload journalctl log files. If this is considered spam, however, please let me know.

Thanks for your time! =]

szb added a comment.Apr 21 2021, 12:23 PM

Two more undesired log-offs occurred, attaching the journalctl files for the last hour before they took place.

I don't know how to interpret most of them, but I can see a lot of error messages in there, some related to budgie, some to gnome-screensaver, ... But they don't seem to happen right before log-out, so I don't know if they are related to the problem or not.

szb added a comment.Apr 22 2021, 9:27 PM

Many more crashes/logouts and therefore many more (hopefully meaningful) logfiles.

It seems that several users experience the very same issue now, regardless of hardware or desktop environment (i.e.: https://discuss.getsol.us/d/6645-system-instabilities-using-openbox/).

Admittedly, I wasn't specific enough. Journalctl only tells me that Xorg is crashing. I need the Xorg logs to know why it is crashing, which is logged separately.

That said, there's no way everyone here is experiencing the same issue as someone was in February or even March. In the future this should be reported separately because a significant amount of the repo has changed since the OP created this task and @palto42 followed up. I agree that you all seem to have a crashing Xorg server, but that's a symptom, not the root cause.

Hope this helps as I think it captured an error. For info I have a i5 processor too.

@bigrammy Can you confirm if you have xorg-driver-video-intel installed? If so, can you test if removing it resolves the issue? Thanks.

DataDrake triaged this task as Needs More Info priority.Apr 22 2021, 10:41 PM
DataDrake edited projects, added Software; removed Lacks Project.
bigrammy added a comment.EditedApr 22 2021, 11:23 PM

@DataDrake EDIT: I removed the driver using the software center and will let you know if the crashes continue.

OS: Solus x86_64
Host: W25CSW
Kernel: 5.11.12-177.current
Uptime: 3 hours, 3 mins
Packages: 961 (eopkg), 11 (flatpak)
Shell: bash 5.1.4
Resolution: 1366x768
DE: Budgie (git-481e9bd25afaa4f35ae9d
Theme: Plata [GTK2/3]
Icons: Papirus [GTK2/3]
Terminal: gnome-terminal
CPU: Intel i5-4210M (4) @ 3.200GHz
GPU: Intel 4th Gen Core Processor
Memory: 1584MiB / 15906MiB

@DataDrake On reboot I was faced with the black screen with no prompt etc so I had to do a chroot from solus live usb and re install the xorg-driver-video-intel. So definitely not a fix LOL unless re installing the driver fixes things. I will let you know if the crashes continue.

@bigrammy That's a weird one, because it should have fallen back to the modesetting driver.

szb added a comment.Apr 23 2021, 4:38 PM

@DataDrake I also tried your suggestion - with the same result, unfortunately. Can't get past my laptop's boot screen now.

@bigrammy Could you kindly tell me in more detail how you restored your system, please? I looked for instructions, but I couldn't find much that seemed applicable, besides a possibly outdated guide for Ubuntu.

bigrammy added a comment.EditedApr 23 2021, 6:47 PM

@szb Sorry for the late reply. I used this guide https://getsol.us/articles/troubleshooting/boot-rescue/en/

Once you have mounted everything and in chroot with a working internet connection add the xorg-driver-video-intel using the terminal.

eopkg install xorg-driver-video-intel Then I did the usysconf run -f and exit after all was done just shut down and restart the PC and hopefully all will be fixed.

@DataDrake It crashed again. This time I woke the laptop from sleeping with a press of the power key which brought me to the login screen as normal so I entered my password and hit enter then black screen again. No flashing cursor no pointer nothing.
After this I have to force power off & boot the Solus Live USB then shutdown properly and reboot to get everything back to normal. EDIT: This has happened 3 times now since posting.
Since szb had the same problem as me when removing the xorg-driver-video-intel it's clearly it not falling back to the modesetting driver so I would remove or edit that post asking to try remove the xorg-driver-video-intel to prevent anyone else having the hassle of trying to re install it again :)
Here are the logs.

szb added a comment.Apr 24 2021, 10:27 AM

@bigrammy Thank you very much! I'm back up and running! :D My bad at didn't directly look at the Solus guide, didn't see the forest for the trees.

@DataDrake Attaching my X.org log as well, in case it helps!

bigrammy added a comment.EditedApr 24 2021, 10:52 AM

@szb Your welcome I am glad you got it sorted.

@DataDrake I have reverted to the old Linux-lts (com.solus-project.lts.4.14.226-170) via the Software Center and I have not had any issues to date.
I am starting to think this maybe a 5.x Kernel issue now.
I will remain on this old Kernel for a while and will report if I have any issues regarding these crashes.
PS: How do we force the modesetting driver as default??

@DataDrake No issues at all with Old kernel so it appears something is incompatible with the newer 5.11 kernel.

Yesterday I had a crash after wake-up from sleep, in journalctl I can see an entry

Apr 24 14:36:01 cubitus audit[742]: ANOM_ABEND auid=4294967295 uid=0 gid=0 ses=4294967295 subj=unconfined pid=742 comm="Xorg" exe="/usr/lib64/xorg-server/Xorg" sig=6 res=1

But the signal at the end of the /var/log/Xorg.0.log.old seems to differ, jounalctl message shows sig=6 (is this referring to term signal?) and log reports Caught signal 11 (Segmentation fault)

[ 11416.222] (II) XINPUT: Adding extended input device "Logitech K360" (type: KEYBOARD, id 15)
[ 11416.222] (**) Option "xkb_model" "pc104"
[ 11416.222] (**) Option "xkb_layout" "de"
[ 11416.226] (II) config/udev: Adding input device Logitech M585/M590 (/dev/input/event5)
[ 11416.226] (**) Logitech M585/M590: Applying InputClass "libinput pointer catchall"
[ 11416.226] (**) Logitech M585/M590: Applying InputClass "libinput keyboard catchall"
[ 11416.226] (**) Logitech M585/M590: Applying InputClass "system-keyboard"
[ 11416.226] (II) Using input driver 'libinput' for 'Logitech M585/M590'
[ 11416.226] (**) Logitech M585/M590: always reports core events
[ 11416.226] (**) Option "Device" "/dev/input/event5"
[ 11416.226] (**) Option "_source" "server/udev"
[ 11416.229] (II) event5  - Logitech M585/M590: is tagged by udev as: Keyboard Mouse
[ 11416.229] (II) event5  - Logitech M585/M590: device is a pointer
[ 11416.229] (II) event5  - Logitech M585/M590: device is a keyboard
[ 11416.229] (II) event5  - Logitech M585/M590: device removed
[ 11416.245] (II) libinput: Logitech M585/M590: needs a virtual subdevice
[ 11416.245] (**) Option "config_info" "udev:/sys/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.1/2-1.1.3/2-1.1.3:1.2/0003:046D:C52B.000C/0003:046D:406B.000D/input/input39/event5"
[ 11416.245] (II) XINPUT: Adding extended input device "Logitech M585/M590" (type: MOUSE, id 19)
[ 11416.246] (**) Option "AccelerationScheme" "none"
[ 11416.248] (**) Logitech M585/M590: (accel) selected scheme none/0
[ 11416.248] (**) Logitech M585/M590: (accel) acceleration factor: 2.000
[ 11416.248] (**) Logitech M585/M590: (accel) acceleration threshold: 4
[ 11416.251] (II) event5  - Logitech M585/M590: is tagged by udev as: Keyboard Mouse
[ 11416.252] (II) event5  - Logitech M585/M590: device is a pointer
[ 11416.252] (II) event5  - Logitech M585/M590: device is a keyboard
[ 11416.254] (**) Logitech M585/M590: Applying InputClass "libinput pointer catchall"
[ 11416.254] (**) Logitech M585/M590: Applying InputClass "libinput keyboard catchall"
[ 11416.254] (**) Logitech M585/M590: Applying InputClass "system-keyboard"
[ 11416.254] (II) Using input driver 'libinput' for 'Logitech M585/M590'
[ 11416.254] (**) Logitech M585/M590: always reports core events
[ 11416.254] (**) Option "Device" "/dev/input/event5"
[ 11416.254] (**) Option "_source" "_driver/libinput"
[ 11416.254] (II) libinput: Logitech M585/M590: is a virtual subdevice
[ 11416.254] (**) Option "config_info" "udev:/sys/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.1/2-1.1.3/2-1.1.3:1.2/0003:046D:C52B.000C/0003:046D:406B.000D/input/input39/event5"
[ 11416.254] (II) XINPUT: Adding extended input device "Logitech M585/M590" (type: KEYBOARD, id 20)
[ 11416.254] (**) Option "xkb_model" "pc104"
[ 11416.254] (**) Option "xkb_layout" "de"
[ 11425.372] (EE)
[ 11425.372] (EE) Backtrace:
[ 11425.375] (EE) 0: /usr/lib64/xorg-server/Xorg (OsLookupColor+0x135) [0x5a5f85]
[ 11425.376] (EE) 1: /usr/lib/libpthread.so.0 (funlockfile+0x60) [0x7fd1bc0ae310]
[ 11425.376] (EE) 2: /usr/lib64/xorg-server/Xorg (RamDacHandleColormaps+0xbbf) [0x4c12df]
[ 11425.377] (EE) 3: /usr/lib64/xorg-server/Xorg (AddTraps+0x5f64) [0x5195f4]
[ 11425.377] (EE) 4: /usr/lib64/xorg-server/Xorg (FreeCursor+0x57) [0x431ce7]
[ 11425.377] (EE) 5: /usr/lib64/xorg-server/Xorg (ConfineToShape+0xa34) [0x448444]
[ 11425.377] (EE) 6: /usr/lib64/xorg-server/Xorg (WindowHasNewCursor+0x37) [0x449907]
[ 11425.378] (EE) 7: /usr/lib64/xorg-server/Xorg (ChangeWindowAttributes+0xd2d) [0x46e0bd]
[ 11425.378] (EE) 8: /usr/lib64/xorg-server/Xorg (ProcBadRequest+0x223) [0x438893]
[ 11425.378] (EE) 9: /usr/lib64/xorg-server/Xorg (SendErrorToClient+0x364) [0x43ed34]
[ 11425.378] (EE) 10: /usr/lib64/xorg-server/Xorg (InitFonts+0x3b4) [0x442f14]
[ 11425.380] (EE) 11: /usr/lib/libc.so.6 (__libc_start_main+0xf2) [0x7fd1bbedba92]
[ 11425.380] (EE) 12: /usr/lib64/xorg-server/Xorg (_start+0x2e) [0x42c02e]
[ 11425.380] (EE)
[ 11425.380] (EE) Segmentation fault at address 0x0
[ 11425.380] (EE)
Fatal server error:
[ 11425.380] (EE) Caught signal 11 (Segmentation fault). Server aborting
[ 11425.380] (EE)
[ 11425.380] (EE)
Please consult the The X.Org Foundation support
         at http://wiki.x.org
 for help.
[ 11425.380] (EE) Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[ 11425.380] (EE)
[ 11425.380] (II) AIGLX: Suspending AIGLX clients for VT switch
[ 11426.123] (EE) Server terminated with error (1). Closing log file.

Some weeks back I tried the LTS kernel, but same issues.

My impression is that in the last weeks the issues are less often (only every 5-10 wake-up's), but not sure if something changed in my usage pattern or it's because of some package update.

Is anything happening or in the pipeline with this issue?
Reading the forum it clearly does not affect everyone.
I am not sure what is causing this issue for me and a few others to varying degrees of severity.
Just to be clear my Laptop does the following.
The system will randomly switch to the log on screen.
All app's and open windows in the session are closed when logging back in.
or It goes to the black screen with no pointer or blinking cursor & I need to do a force power off and recover as described in the previous posts.

  1. Issues seems to of started two kernel versions ago with the latest causing most crashes.
  1. Old LTS Kernel in the Software Center has no issues.

@JoshStrobl @DataDrake

Thanks in Advance.

We don't have enough information to act on this. So far all I've seen is that a subset of Intel CPUs are crashing Xorg under some unknown circumstance. And that could just as easily be tied to the xorg-server update.

I was consistently getting 1-2 spontaneous log-offs immediately after logging in after waking, or after the screen had locked. Something has changed with the last day or so of updates in unstable. I no longer see the behavior. Hardware: Latitude Laptop, CPU Intel i5-4310U, GPU Intel Haswell-ULT

Is there anything we can do to produce more detailed or helpful logs for you ?

I found this https://www.linuxfromscratch.org/blfs/view/svn/x/x7driver.html#xorg-intel-driver
I did noticed this remark

Working around problems with the Intel driver
The SandyBridge New Acceleration (SNA) code is intended to replace the old UXA (UMA Acceleration Architecture), but it is a large body of code and may cause problems. However, the version of the code in the version control system listed above has been tested successfully with both SNA and UXA capabilities.
To work around this problem, as well as enabling support for UXA, it is necessary to force UXA to be used by creating a configuration file. If this problem applies to you, create the following file as the root user and modify it as needed:

cat >> /etc/X11/xorg.conf.d/20-intel.conf << "EOF"
Section   "Device"
        Identifier "Intel Graphics"
        Driver     "intel"
        #Option     "DRI" "2"            # DRI3 is default
        #Option     "AccelMethod"  "sna" # default
        #Option     "AccelMethod"  "uxa" # fallback
EndSection
EOF

So my question is was our driver version compiled using the --enable-uxa flag to allow setting uxa mode via a user created /etc/X11/xorg.conf.d/20-intel.conf ?

That's intended as a workaround for much older systems if it is forcing DRI2, so I'd be surprised if that fixed anything for you. But yes, that flag is set during the build process.

The Latest Solus OS update using the latest kernel com.solus-project.current.5.11.16-178 seems to of fixed things me. : )
No crashes at all since the update.
Thanks to those responsible for fixing things.

I still face issues with spontaneous log-offs, but quite rarely and not always immediately after wake-up from sleep.
Too rare to spend more time in debugging in my view - just hope that it fully disappears at some point :)