Desktop frozen for 30 seconds after Nvidia driver crash

[Mathieu Tarral]

Hi,

I just experienced a frozen desktop, where I could barely move the mouse, for maybe 30 seconds or more.

When the desktop recovered, I looked at dmesg logs to find these lines:

[ 461.937702] NVRM: GPU at PCI:0000:01:00: GPU-734060ac-116d-fbbb-1553-d84896b846a4
[ 461.937710] NVRM: Xid (PCI:0000:01:00): 31, pid=275, Ch 0000002e, intr 00000000. MMU Fault: ENGINE HOST0 HUBCLIENT_HOST_CPU faulted @ 0x1_0036d000. Fault is of type FAULT_PTE ACCESS_TYPE_VIRT_READ

It looks like a fault from the Nvidia driver.
I'm using the tested one "nvidia-driver-435", installed via the GUI tool.

Could you report this to Nvidia please ?

Thanks

ProblemType: Bug
DistroRelease: Ubuntu 19.10
Package: nvidia-driver-435 435.21-0ubuntu2
ProcVersionSignature: Ubuntu 5.3.0-19.20-generic 5.3.1
Uname: Linux 5.3.0-19-generic x86_64
NonfreeKernelModules: nvidia_modeset nvidia
ApportVersion: 2.20.11-0ubuntu8.2
Architecture: amd64
CurrentDesktop: ubuntu:GNOME
Date: Sun Nov 10 17:35:12 2019
InstallationDate: Installed on 2019-09-26 (45 days ago)
InstallationMedia: Ubuntu 19.04 "Disco Dingo" - Release amd64 (20190416)
SourcePackage: nvidia-graphics-drivers-435
UpgradeStatus: No upgrade log present (probably fresh install)

[Fabio Ornellas (fabio-ornellas)]

I'm having the same issue with Ubuntu Focal, kernel 5.4.0-40-generic and latest Nvidia 450.57 from https://launchpad.net/~graphics-drivers/+archive/ubuntu/ppa.

I started having this issue when I update Chrome **after** 81.0.4044.138. Up to this version, things work. From this Chrome version on, whenever I open any OpenGL site (eg: onshape.com), everything freezes for a good 30s, I get a message on "OpenGL not working" from the website, and kernel spits this:

[ 80.762915] NVRM: GPU at PCI:0000:09:00: GPU-824bae61-b6cb-f859-5c4b-531553afa61b
[ 80.762918] NVRM: GPU Board Serial Number:
[ 80.762922] NVRM: Xid (PCI:0000:09:00): 31, pid=5024, Ch 00000016, intr 00000000. MMU Fault: ENGINE HOST0 HUBCLIENT_HOST_CPU faulted @ 0x1_01a8e000. Fault is of type FAULT_PTE ACCESS_TYPE_VIRT_READ

[Fabio Ornellas (fabio-ornellas)]

PS: just tested with latest vanilla kernel 5.7.9-050709-generic from https://kernel.ubuntu.com/~kernel-ppa/mainline/?C=N;O=D, things work. It seems there's some issue with 5.4.0-40-generic.

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Albert Zeyer (albertzeyer)]

I think I have a very related (or the same) error (or multiple errors).
A lot of logs here: https://askubuntu.com/questions/1236721/desktop-hung-up-freeze-gpuwatchdog-segfault-nvidia-frontend-close

(Not sure if they are all the same cause, or different... I think they are the same.)

When you search the Internet for "NVRM: Xid 31" errors, you will find quite a lot of results, also a lot of recent results.

In my (most recent) case, I have Nvidia driver 450.80.02, and Linux kernel 5.4.0-59.

[jasoncollege24 (jasoncollege24)]

I may also have the same, or related issue.

OS: Kubuntu 20.04 LTS
CPU: Intel Core i9-9900K
GPU: GeForce RTX 2080 SUPER
Driver: Nvidia proprietary driver 460.32.03
Linux Kernel: 5.4.0-65-generic

What happened?
Was watching a movie on Plex, using the web player in the Brave browser (List as Version 1.19.88 Chromium: 88.0.4324.96 (Official Build) (64-bit)), when the system froze. I could hear sound for several seconds in the background, then it just stopped completely.

What was expected?
Should've continued playing the video.

Log:
NVRM: Xid (PCI:0000:01:00): 31, pid=278, Ch 00000002, intr 00000000. MMU Fault: ENGINE HOST0 HUBCLIENT_HOST faulted @ 0x10_00000000. Fault is of type FAULT_PDE ACCESS_TYPE_VIRT_READ
NVRM: Xid (PCI:0000:01:00): 31, pid=5096, Ch 00000050, intr 00000000. MMU Fault: ENGINE HOST0 HUBCLIENT_HOST faulted @ 0x21_02b07000. Fault is of type FAULT_PDE ACCESS_TYPE_VIRT_READ