[nvidia] Screen frozen after switch to other tty and do login & logout

Bug #1821194 reported by Vic Liu
20
This bug affects 1 person
Affects Status Importance Assigned to Milestone
OEM Priority Project
New
Undecided
Cyrus Lien
gnome-shell (Ubuntu)
New
Undecided
Unassigned
linux (Ubuntu)
Incomplete
Undecided
Unassigned
nvidia-graphics-drivers-390 (Ubuntu)
New
Undecided
Unassigned
nvidia-graphics-drivers-410 (Ubuntu)
New
Undecided
Unassigned

Bug Description

To simplify reproduce steps, one only has to:
1. install (tested so far) Ubuntu Bionic
2. switch to other tty by pressing Ctrl+Alt+Fn
3. login and logout
Then it should try to switch back to GDM (no matter you have logged in GUI or not), triggers vt-switch bug and lock up.

Some more findings:

1. So far we know all other newer Ubuntu versions are immune, e.g. Ubuntu Cosmic's gdm(version 3.30.1-1ubuntu5).

2. login _and_ logout from another virtual console is required. Switching between text console and GUI back and forth before logging out doesn't trigger this bug. So this might(?) also involves systemd-logind as well;

3. while gdm locked up, kernel and the rest of the system are actually still alive. Networking is still available, and even the input devices are still functioning well. Only gdm fails to grab the device inputs.

4. when gdm locks up due to either resume from suspend or vt-switch, restart gdm (from remote ssh) will bring the system back.

### Disco
| |4.15|4.18|4.19|5.0
|---|----|----|----|---
|nouveau| | | |
|nvidia-390 bionic|Affected|Affected|Affected|Affected
|nvidia-390 cosmic|Affected|Affected|Affected|Affected
|nvidia-410|Affected|Affected|Affected|Affected
Affected: only gdm(vt-1) fails to grab further input. Use `chvt` to switch to other vt in a remote shell.

#### Packages:
* kernel:
   * 4.15.0-46-generic
   * 4.18.0-16-generic
   * 4.19.0-13-generic
   * 5.0.0-7-generic
* nvidia-driver:
   * 390.116-0ubuntu0.18.04.1
   * 390.116-0ubuntu0.18.10.1
   * 410.104-0ubuntu1
* systemd:
   * 240-6ubuntu1
* gdm3:
   * 3.31.4+git20190225-1ubuntu1
#### Hardware
* 201811-26623

Summary: in Ubuntu Disco, nouveau driver is immune from this issue in all kernel versions tested. In contrary, nVIDIA driver fails every case.

 ### Cosmic
| |4.15|4.18|4.19|5.0
|-----------------|----|----|----|---
|nouveau | | | |
|nvidia-390 bionic|Affected|Affected|Affected|Affected
|nvidia-390 cosmic|Affected|Affected|Affected|Affected
|nvidia-410 |Affected|Affected|Affected|Affected
Affected: only gdm(vt-1) fails to grab further input. Use `chvt` to switch to other vt in a remote shell.
#### Packages:
* kernel:
   * 4.15.0-46-generic
   * 4.18.0-16-generic
   * 4.19.0-13-generic
   * 5.0.0-7-generic
* nvidia-driver:
   * 390.116-0ubuntu0.18.04.1
   * 390.116-0ubuntu0.18.10.1
   * 410.104-0ubuntu1
* systemd:
   * 239-7ubuntu10
* gdm3:
   * 3.30.1-1ubuntu5
#### Hardware
* 201811-26623

Summary: in Ubuntu Cosmic, same with Disco.

# Bionic
| |4.15|4.18 |4.19
|-----------------|----|------------|----
|nouveau|Affected[1]| |Affected[1]
|hwe xorg|Affected[1]| |Affected[1]
|nvidia-390 bionic+hwe xorg|Affected[1]|Affected[1]|Affected[1]
|nvidia-390 bionic|Affected[2]|Affected[1]|Affected[1]
|nvidia-390 cosmic|Affected[1]|Affected[1]|Affected[1]
|nvidia-410 |Affected[2]|Affected[1]|Affected[3]
[1]: no vt-switch inside gnome-shell. Press Ctrl+Alt+Fn in gdm instead. Only gdm(vt-1) fails to grab further input. Use `chvt` to switch to other vt in a remote shell.
[2]: vt-switch inside gnome-shell causes lockup. Cannot switch to vt in gdm (but no system lockup). Only gdm(vt-1) fails to grab further input. Use `chvt` to switch to other vt in a remote shell.
[3]: vt-switch inside gnome-shell causes lockup. Press Ctrl+Alt+Fn in gdm instead. Only gdm(vt-1) fails to grab further input. Use `chvt` to switch to other vt in a remote shell.
## Packages:
* kernel:
   * 4.15.0-46-generic
   * 4.18.0-16-generic
   * 4.19.0-13-generic
* nvidia-driver:
   * 390.116-0ubuntu0.18.04.1
   * 390.116-0ubuntu0.18.10.1
   * 410.104-0ubuntu1
* systemd:
   * 237-3ubuntu10.12
* gdm3:
   * 3.28.3-0ubuntu18.04.3
## Hardware
* 201811-26623

Summary: only 4.18 kernel with nouveau driver is immune from this issue, which follows some changes in user land should be necessary to stabilize all the (nouveau) cases. But even with that, from previous results of Cosmic/Disco, there should still some works to be done in the nvidia driver.

OEM-tag: LP: #1804607

Revision history for this message
Daniel van Vugt (vanvugt) wrote :

If you can still ssh in while the screen is frozen then please do so and run:

  dmesg > frozendmesg.txt

and send us the file 'frozendmesg.txt'

summary: - Screen frozen after switch to other tty and do login & logout
+ [nvidia] Screen frozen after switch to other tty and do login & logout
affects: gdm3 (Ubuntu) → nvidia-graphics-drivers-390 (Ubuntu)
tags: added: nvidia
tags: added: bionic cosmic
Changed in linux (Ubuntu):
status: New → Incomplete
Changed in nvidia-graphics-drivers-390 (Ubuntu):
status: New → Incomplete
Changed in nvidia-graphics-drivers-410 (Ubuntu):
status: New → Incomplete
Revision history for this message
Cyrus Lien (cyruslien) wrote :
Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Thanks. It looks like the last kernel message is one from the Nvidia driver, but I can't tell if it's an error or just informational:

[ 871.681257] NVRM: GPU at PCI:0000:01:00: GPU-7eba71ac-fd71-c5c9-f569-09f6406e66c7
[ 871.681262] NVRM: GPU Board Serial Number:
[ 871.681264] NVRM: Xid (PCI:0000:01:00): 16, Head 00000000 Count 00008a93

Revision history for this message
Daniel van Vugt (vanvugt) wrote :
Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Again, if you can still ssh in while the screen is frozen then please do so and run:

  journalctl -b0 > frozenjournal.txt

and send us the file 'frozenjournal.txt'

Revision history for this message
Cyrus Lien (cyruslien) wrote :

The directory /var/crash is empty.

Revision history for this message
Cyrus Lien (cyruslien) wrote : apport information

ProblemType: Bug
ApportVersion: 2.20.9-0ubuntu7.2
Architecture: amd64
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC1: u 3364 F.... pulseaudio
 /dev/snd/controlC0: u 3364 F.... pulseaudio
DistributionChannelDescriptor:
 # This is the distribution channel descriptor for the OEM CDs
 # For more information see http://wiki.ubuntu.com/DistributionChannelDescriptor
 canonical-oem-somerville-bionic-amd64-20180608-47+beaver-gambit-eagle-kylin+X45
DistroRelease: Ubuntu 18.04
InstallationDate: Installed on 2019-03-22 (0 days ago)
InstallationMedia: Ubuntu 18.04 "Bionic" - Build amd64 LIVE Binary 20180608-09:38
MachineType: Dell Inc. Inspiron 3670
NonfreeKernelModules: nvidia_modeset nvidia
Package: nvidia-graphics-drivers-410
ProcEnviron:
 TERM=xterm-256color
 PATH=(custom, no user)
 XDG_RUNTIME_DIR=<set>
 LANG=en_US.UTF-8
 SHELL=/bin/bash
ProcFB: 0 EFI VGA
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-4.15.0-1029-oem.efi.signed root=UUID=baab0cd6-ff41-4a5b-93a0-c52386e24070 ro mem_sleep_default=deep quiet splash vt.handoff=1
ProcVersionSignature: Ubuntu 4.15.0-1029.34-oem 4.15.18
RelatedPackageVersions:
 linux-restricted-modules-4.15.0-1029-oem N/A
 linux-backports-modules-4.15.0-1029-oem N/A
 linux-firmware 1.173.2
Tags: bionic
Uname: Linux 4.15.0-1029-oem x86_64
UpgradeStatus: No upgrade log present (probably fresh install)
UserGroups: adm cdrom dip lpadmin plugdev sambashare sudo
_MarkForUpload: True
dmi.bios.date: 03/04/2019
dmi.bios.vendor: Dell Inc.
dmi.bios.version: 2.7.2
dmi.board.vendor: Dell Inc.
dmi.chassis.type: 3
dmi.chassis.vendor: Dell Inc.
dmi.modalias: dmi:bvnDellInc.:bvr2.7.2:bd03/04/2019:svnDellInc.:pnInspiron3670:pvr:rvnDellInc.:rn:rvr:cvnDellInc.:ct3:cvr:
dmi.product.family: Inspiron
dmi.product.name: Inspiron 3670
dmi.sys.vendor: Dell Inc.

tags: added: apport-collected
Revision history for this message
Cyrus Lien (cyruslien) wrote : AlsaInfo.txt

apport information

Revision history for this message
Cyrus Lien (cyruslien) wrote : CRDA.txt

apport information

Revision history for this message
Cyrus Lien (cyruslien) wrote : CurrentDmesg.txt

apport information

Revision history for this message
Cyrus Lien (cyruslien) wrote : IwConfig.txt

apport information

Revision history for this message
Cyrus Lien (cyruslien) wrote : Lspci.txt

apport information

Revision history for this message
Cyrus Lien (cyruslien) wrote : Lsusb.txt

apport information

Revision history for this message
Cyrus Lien (cyruslien) wrote : ProcCpuinfo.txt

apport information

Revision history for this message
Cyrus Lien (cyruslien) wrote : ProcCpuinfoMinimal.txt

apport information

Revision history for this message
Cyrus Lien (cyruslien) wrote : ProcInterrupts.txt

apport information

Revision history for this message
Cyrus Lien (cyruslien) wrote : ProcModules.txt

apport information

Revision history for this message
Cyrus Lien (cyruslien) wrote : PulseList.txt

apport information

Revision history for this message
Cyrus Lien (cyruslien) wrote : RfKill.txt

apport information

Revision history for this message
Cyrus Lien (cyruslien) wrote : UdevDb.txt

apport information

Revision history for this message
Cyrus Lien (cyruslien) wrote : WifiSyslog.txt

apport information

Revision history for this message
Daniel van Vugt (vanvugt) wrote :

I'm really not sure what's going on there. There's no evidence of a crash and if there is a hang then this looks like the last things gnome-shell said:

Mar 22 04:07:03 u-Inspiron-3670 gnome-shell[1577]: g_array_unref: assertion 'array' failed
Mar 22 04:07:03 u-Inspiron-3670 gnome-shell[1577]: g_array_unref: assertion 'array' failed
Mar 22 04:07:03 u-Inspiron-3670 gnome-shell[1577]: g_array_unref: assertion 'array' failed
Mar 22 04:07:03 u-Inspiron-3670 gnome-shell[1577]: clutter_input_device_get_device_id: assertion 'CLUTTER_IS_INPUT_DEVICE (device)' failed
Mar 22 04:07:03 u-Inspiron-3670 gnome-shell[1577]: clutter_input_device_get_device_id: assertion 'CLUTTER_IS_INPUT_DEVICE (device)' failed
Mar 22 04:07:03 u-Inspiron-3670 gnome-shell[1577]: clutter_input_device_get_device_id: assertion 'CLUTTER_IS_INPUT_DEVICE (device)' failed
Mar 22 04:07:03 u-Inspiron-3670 gnome-shell[1577]: clutter_input_device_get_device_id: assertion 'CLUTTER_IS_INPUT_DEVICE (device)' failed
Mar 22 04:07:03 u-Inspiron-3670 gnome-shell[1577]: clutter_input_device_get_device_id: assertion 'CLUTTER_IS_INPUT_DEVICE (device)' failed
Mar 22 04:07:03 u-Inspiron-3670 gnome-shell[1577]: clutter_input_device_get_device_id: assertion 'CLUTTER_IS_INPUT_DEVICE (device)' failed
Mar 22 04:07:03 u-Inspiron-3670 gnome-shell[1577]: clutter_input_device_get_device_id: assertion 'CLUTTER_IS_INPUT_DEVICE (device)' failed
Mar 22 04:07:03 u-Inspiron-3670 gnome-shell[1577]: clutter_input_device_get_device_id: assertion 'CLUTTER_IS_INPUT_DEVICE (device)' failed
Mar 22 04:07:03 u-Inspiron-3670 gnome-shell[1577]: clutter_input_device_get_device_id: assertion 'CLUTTER_IS_INPUT_DEVICE (device)' failed
Mar 22 04:07:11 u-Inspiron-3670 kernel: NVRM: GPU at PCI:0000:01:00: GPU-7eba71ac-fd71-c5c9-f569-09f6406e66c7
Mar 22 04:07:11 u-Inspiron-3670 kernel: NVRM: GPU Board Serial Number:
Mar 22 04:07:11 u-Inspiron-3670 kernel: NVRM: Xid (PCI:0000:01:00): 16, Head 00000000 Count 00008a93
Mar 22 04:07:11 u-Inspiron-3670 gnome-shell[1577]: clutter_input_device_get_device_type: assertion 'CLUTTER_IS_INPUT_DEVICE (device)' failed
Mar 22 04:07:11 u-Inspiron-3670 gnome-shell[1577]: JS ERROR: TypeError: device is null
                                                   _init/<@resource:///org/gnome/shell/ui/keyboard.js:518:1

It's slightly interesting the kernel message from Nvidia is in the middle.

Changed in linux (Ubuntu):
status: Incomplete → New
Changed in nvidia-graphics-drivers-390 (Ubuntu):
status: Incomplete → New
Changed in nvidia-graphics-drivers-410 (Ubuntu):
status: Incomplete → New
You-Sheng Yang (vicamo)
tags: added: disco
Revision history for this message
Ubuntu Kernel Bot (ubuntu-kernel-bot) wrote : Missing required logs.

This bug is missing log files that will aid in diagnosing the problem. While running an Ubuntu kernel (not a mainline or third-party kernel) please enter the following command in a terminal window:

apport-collect 1821194

and then change the status of the bug to 'Confirmed'.

If, due to the nature of the issue you have encountered, you are unable to run this command, please add a comment stating that fact and change the bug status to 'Confirmed'.

This change has been made by an automated script, maintained by the Ubuntu Kernel Team.

Changed in linux (Ubuntu):
status: New → Incomplete
Alex Tu (alextu)
Changed in oem-priority:
assignee: nobody → Cyrus Lien (cyruslien)
Revision history for this message
You-Sheng Yang (vicamo) wrote :

This was previously discovered as an alternative way to trigger bug 1804607, but re-filed as a new one for the different behaviour.

By the way, bug 1817738 was once considered a potential root cause to it. However the proposed change prevents gnome-shell from starting up normally on systems with nvidia cards and installed nvidia-drivers.

Vic Liu (zongminl)
description: updated
Revision history for this message
You-Sheng Yang (vicamo) wrote :

Updates from bug 1804607, this is a known xorg issue and duplicates bug 1822184 that can be worked around by disable clear_console in the $HOME/.bash_logout:

  $ cat $HOME/.bash_logout
  # ~/.bash_logout: executed by bash(1) when login shell exits.

  # when leaving the console clear the screen to increase privacy
  if [ "$SHLVL" = 1 ]; then
    # [ -x /usr/bin/clear_console ] && /usr/bin/clear_console -q
  fi

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.