Can not login after new kernel installed (nvidia driver crashed)

[vodopad27]

Hello. Today i have installed new kernel from. I installed 5.15.0-50 and after that i can not login to system... I see boot text, but after booting i see black log screen and can not login. I able boot and login with old kernel (5.15.0-48).

I updated these packages:
Install: linux-headers-5.15.0-50-generic:amd64 (5.15.0-50.56, automatic), linux-cloud-tools-5.15.0-50-generic:amd64 (5.15.0-50.56, automatic), linux-modules-5.15.0-50-generic:amd64 (5.15.0-50.56, automatic), linux-modules-extra-5.15.0-50-generic:amd64 (5.15.0-50.56, automatic), linux-cloud-tools-5.15.0-50:amd64 (5.15.0-50.56, automatic), linux-tools-5.15.0-50:amd64 (5.15.0-50.56, automatic), linux-headers-5.15.0-50:amd64 (5.15.0-50.56, automatic), linux-image-5.15.0-50-generic:amd64 (5.15.0-50.56, automatic), linux-tools-5.15.0-50-generic:amd64 (5.15.0-50.56, automatic)
Upgrade: linux-headers-generic:amd64 (5.15.0.48.48, 5.15.0.50.50), google-chrome-stable:amd64 (106.0.5249.103-1, 106.0.5249.119-1), linux-generic:amd64 (5.15.0.48.48, 5.15.0.50.50), linux-image-generic:amd64 (5.15.0.48.48, 5.15.0.50.50), linux-tools-generic:amd64 (5.15.0.48.48, 5.15.0.50.50), linux-cloud-tools-generic:amd64 (5.15.0.48.48, 5.15.0.50.50)

Updated logs see in attachement.

Logs from journalctl see in attachements.

ProblemType: Bug
DistroRelease: Ubuntu 22.04
Package: linux-image-5.15.0-50-generic 5.15.0-50.56
ProcVersionSignature: Ubuntu 5.15.0-48.54-generic 5.15.53
Uname: Linux 5.15.0-48-generic x86_64
NonfreeKernelModules: nvidia_modeset nvidia
ApportVersion: 2.20.11-0ubuntu82.1
Architecture: amd64
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC0: vodka 1819 F.... pulseaudio
 /dev/snd/controlC1: vodka 1819 F.... pulseaudio
CasperMD5CheckResult: unknown
Date: Wed Oct 12 19:56:43 2022
InstallationDate: Installed on 2018-05-02 (1623 days ago)
InstallationMedia: Ubuntu 18.04 LTS "Bionic Beaver" - Release amd64 (20180426)
Lsusb:
 Bus 002 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
 Bus 001 Device 004: ID 05e3:0608 Genesys Logic, Inc. Hub
 Bus 001 Device 003: ID 09da:fa18 A4Tech Co., Ltd. USB Device
 Bus 001 Device 002: ID 046d:c084 Logitech, Inc. G203 Gaming Mouse
 Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
MachineType: Gigabyte Technology Co., Ltd. H610M S2H DDR4
ProcFB: 0 VESA VGA
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-5.15.0-48-generic root=UUID=b27e8e45-cedd-4ab6-b2e6-ab6bef5e9336 ro
PulseList: Error: command ['pacmd', 'list'] failed with exit code 1: No PulseAudio daemon running, or not running as session daemon.
RelatedPackageVersions:
 linux-restricted-modules-5.15.0-48-generic N/A
 linux-backports-modules-5.15.0-48-generic N/A
 linux-firmware 20220329.git681281e4-0ubuntu3.5
RfKill:

SourcePackage: linux
UpgradeStatus: Upgraded to jammy on 2022-04-27 (167 days ago)
dmi.bios.date: 03/28/2022
dmi.bios.release: 5.24
dmi.bios.vendor: American Megatrends International, LLC.
dmi.bios.version: F7a
dmi.board.asset.tag: Default string
dmi.board.name: H610M S2H DDR4
dmi.board.vendor: Gigabyte Technology Co., Ltd.
dmi.board.version: x.x
dmi.chassis.asset.tag: Default string
dmi.chassis.type: 3
dmi.chassis.vendor: Default string
dmi.chassis.version: Default string
dmi.modalias: dmi:bvnAmericanMegatrendsInternational,LLC.:bvrF7a:bd03/28/2022:br5.24:svnGigabyteTechnologyCo.,Ltd.:pnH610MS2HDDR4:pvr-CF:rvnGigabyteTechnologyCo.,Ltd.:rnH610MS2HDDR4:rvrx.x:cvnDefaultstring:ct3:cvrDefaultstring:skuDefaultstring:
dmi.product.family: H610 MB
dmi.product.name: H610M S2H DDR4
dmi.product.sku: Default string
dmi.product.version: -CF
dmi.sys.vendor: Gigabyte Technology Co., Ltd.

[vodopad27 (family-gan)]

journalctl log

[vodopad27 (family-gan)]

apt log

[vodopad27 (family-gan)]

Nvidia driver version: 510.85.02
Videocard: RTX 3060

[vodopad27 (family-gan)]

And look to strange line in button after booting...

[vodopad27 (family-gan)]

https://youtu.be/5KiSmPbkibg video with bugs (~1 min)

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

This change was made by a bot.

[Daniel van Vugt (vanvugt)]

Thanks for the bug report. journal.log is showing a crash in the nvidia driver / gpu-manager when using kernel 5.15.0-50

Please try using the 'Additional Drivers' app to install a newer version of the Nvidia driver (like 515).

[vodopad27 (family-gan)]

I have installed 515.65.01, but problem does not resolve. Same problem.

[vodopad27 (family-gan)]

Logs with 515.65.01:

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

This change was made by a bot.

[Amarpreet Singh (solamarpreet)]

I can confirm the bug. My system uses the Nvidia 1050ti chipset. I can only boot the kernel if I set nomodeset parameter in the boot options or using the older 5.15.0-48 kernel

[Launchpad Janitor (janitor)]

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

[Drake Siard (drakesiard)]

Can confirm the bug, with linux-image-5.1.5.50-generic and nvidia-driver-515.

[Alex Shenfield (alex-shenfield)]

Can also confirm this bug using linux-image-5.1.5.50-generic and nvidia-driver-515 and an Nvidia gtx 1070.

All was working under linux-image-5.1.5.48 😭

[Allen (ccwtech)]

Confirmed for me as well.

[Brendan McGrath (redmcg)]

Can also confirm this bug. I'm using a GTX 1050 Ti. Tested (and working) workarounds are:
- use 'nomodeset' boot option
- downgrade to linux-image-5.15.0-48-generic
- use nouveau driver

I tried the following NVIDIA drivers, but they all have this problem (with 5.15.0-50-generic):
- nvidia-driver-515
- nvidia-driver-510
- nvidia-driver-470

[vodopad27 (family-gan)]

Also i tried NVIDIA-Linux-x86_64-520.56.06.run, does not help :-(

[Tom (jimmington)]

Exactly the same issue.

5.15.0-50 did not boot at all.

Went back to 5.15.0-48, Ubuntu software manager claimed some updates had to be done and I did them.

Now 5.15.0-48 no longer boots, my machine is unusable... Not ideal at all...

[Shaofan Qi (shaofan)]

Confirm this.

And why the change log of the kernel package all goes like 'Bump ABI xxx'?

[Peter (pitt0x5f64)]

Same problem, Ubuntu freezes at startup before user login.
- Kernel version: 5.15.0-50-generic
- Nvidia driver version: 515.76
- Nvidia card: RTX3060

[Victor Antonio (victor-antonio-magri)]

Same issue with me.
- Kernel: 5.15.0-50-generic (48 also fails)
- Driver: 515.76-0ubuntu0.22.04.1 amd64
- Card: RTX3060 LHR (GA106).

Currently using the following workaround:
- Kernel: 5.15.0-48-generic
- Driver: 510.85.02-0ubuntu0.22.04.1 amd64

[Richard Travis Crooks (tcrooks0904)]

Same issue for me.
- Kernel: 5.15.0.50-generic
- Driver: 515.65.01
- Card: GTX 1660

Workarounds
1. Kernel 5.15.0.48-generic
2. GRUB "nomodeset"

[Mathieu letombe (mletombe)]

50 and 48 out of order, no workaround works. My pc is dead, i'm writing from my phone.

[Mathieu letombe (mletombe)]

It's working when my TV is off when I power the PC on !!
I was waiting for good news on this page before using my PC again but my cat did power my PC on. I then powered the TV on to see and: everything was working fine with 5.15.0.50-generic and nvidia-driver-515. Nomodeset was not an argument.
I tried to reboot with the TV still on: black screen.
I tried to power the PC on with the TV off: it works!

[Dcintes (dcintes)]

Same problem. Kernel 5.15.0.50-generic, nvidia driver 515, GTX 1660

Workarounds: use nouveau

[vodopad27 (family-gan)]

Same on 5.15.0-52.

[htaccess (linux-gamma)]

I have this issue with 5.15.0.50-generic and 5.15.0.52-generic I could boot with 5.15.0.48-generic and 5.15.0.47-generic

For me some nvidia-515 drivers were held back, I tried the following command:

sudo apt install --upgrade linux-modules-nvidia-515-generic

And this caused further weirdness:

47 => removed
48 => boot hangs on `i915 enabling device (0006 -> 0007)`
50 => can boot if I enter grub and remove quiet and splash from the boot options, but it does not load the nvidia drivers so resolution is terrible
52 => boot hangs with flashing cursor top left ... but if I boot into recovery mode and then resume boot it boots normally including loading the nvidia driver!

After logging in via the recover mode work around `apt update` says `All packages are up to date.` and upgrade wants to autoremove 5.15.0.48-generic

$ nvidia-smi | grep NVIDIA-SMI
| NVIDIA-SMI 515.76 Driver Version: 515.76 CUDA Version: 11.7 |

Looking at the installed nvidia packages I seem to be missing these two which probably explains why 50 does not load the nvidia drivers:

* linux-objects-nvidia-515-5.15.0-50-generic
* linux-modules-nvidia-515-5.15.0-50-generic

The following all seem to be the same issue as far as I can tell:

* https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1993004
* https://www.reddit.com/r/Ubuntu/comments/y0c1k9/cant_boot_on_515050generic_and_nvidia_driver/
* https://askubuntu.com/questions/1435759/nvidia-driver-is-no-longer-working-with-new-kernel
* https://askubuntu.com/questions/1436113/issues-with-linux-5-15-0-50-generic-kubuntu-22-04

[Alberto Milone (albertomilone)]

Can you follow these steps, please?

Remove any nvidia package left on your system:

sudo apt-get --purge remove '*nvidia*'

Finally, install the driver using the following command:

sudo ubuntu-drivers install

Finally, reboot, and see if you can reproduce the problem.

[Franck971 (franck-lefebure)]

Same freeze for me with 515 nv driver.
I'm actually sticking to 5.15.0.48 as a workaround

[htaccess (linux-gamma)]

@Alberto Milone (albertomilone):

sudo apt-get --purge remove '*nvidia*'

removed a lot of packages except these:

```
$ dpkg -l | awk '/nvidia/{ print $2 }'
libnvidia-compute-515:i386
libnvidia-decode-515:i386
libnvidia-encode-515:i386
libnvidia-fbc1-515:i386
```

```
$ sudo ubuntu-drivers install
Traceback (most recent call last):
  File "/usr/bin/ubuntu-drivers", line 513, in <module>
    greet()
  File "/usr/lib/python3/dist-packages/click/core.py", line 1128, in __call__
    return self.main(*args, **kwargs)
  File "/usr/lib/python3/dist-packages/click/core.py", line 1053, in main
    rv = self.invoke(ctx)
  File "/usr/lib/python3/dist-packages/click/core.py", line 1659, in invoke
    return _process_result(sub_ctx.command.invoke(sub_ctx))
  File "/usr/lib/python3/dist-packages/click/core.py", line 1395, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/usr/lib/python3/dist-packages/click/core.py", line 754, in invoke
    return __callback(*args, **kwargs)
  File "/usr/lib/python3/dist-packages/click/decorators.py", line 84, in new_func
    return ctx.invoke(f, obj, *args, **kwargs)
  File "/usr/lib/python3/dist-packages/click/core.py", line 754, in invoke
    return __callback(*args, **kwargs)
  File "/usr/bin/ubuntu-drivers", line 413, in install
    command_install(config)
  File "/usr/bin/ubuntu-drivers", line 187, in command_install
    UbuntuDrivers.detect.nvidia_desktop_pre_installation_hook(to_install)
  File "/usr/lib/python3/dist-packages/UbuntuDrivers/detect.py", line 839, in nvidia_desktop_pre_installation_hook
    with_nvidia_kms = version >= 470
UnboundLocalError: local variable 'version' referenced before assignment
```

looks like this try is failing with a ValueError so the version var is not assigned:

```
$ grep 'Enable KMS if nvidia >= 470' -A 8 /usr/lib/python3/dist-packages/UbuntuDrivers/detect.py
    # Enable KMS if nvidia >= 470
    for package_name in to_install:
        if package_name.startswith('nvidia-driver-'):
            try:
                version = int(package_name.split('-')[-1])
            except ValueError:
                pass
            finally:
                with_nvidia_kms = version >= 470
```

[htaccess (linux-gamma)]

@Alberto Milone (albertomilone):

I went with `sudo apt install nvidia-driver-515` instead due to the `ubuntu-drivers` error above, this installed a bunch of stuff play this lot:

```
Loading new nvidia-515.76 DKMS files...
Building for 5.15.0-52-generic
Building for architecture x86_64
Building initial module for 5.15.0-52-generic
```

rebooted and everything worked.

Looks like I'm back in action (although I assume the two older kernels wont work, hopefully I don't need them).

Thanks for the help!

BTW launchpad formatting is terrible!

[Alberto Milone (albertomilone)]

@htaccess that looks like a bug in ubuntu-drivers. I wonder what the actual value of package_name was though.

[jeremyszu (os369510)]

something like "nvidia-driver-515-open" and which should be fixed by

commit a36807ad1c738abe7980dd09d001549a224b71b3
Author: Alberto Milone <email address hidden>
Date: Tue Aug 23 17:03:01 2022 +0200

    UbuntuDrivers/detect.py: fix driver series detection in nvidia_desktop_pre_installation_hook

Seems like u-d-c from -proposed should fix comment#35.

root cause could be n-g-d-515 shouldn't be landed in -updates before u-d-c.
please see https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-515/+bug/1988836

If someone can enable -proposed and then upgrading u-d-c then try comment#33, then it'll help.

[Dominique Fablauty (download64000)]

Got the same error since updating to 5.15.0-50-generic, and 5.15.0-52-generic doesn't solve the issue.
I have:
- Kernel version: 5.15.0-50-generic or 5.15.0-52-generic
- Driver version: 515.76-0ubuntu0.22.04.1 amd64
- Card: RTX3060

[vodopad27 (family-gan)]

Same on Ubuntu 22.10.

Please. add kinetic to affected distros. I did not find this functionality (probably i do not have rights)

[vodopad27 (family-gan)]

@albertomilone not helped:
sudo apt-get --purge remove '*nvidia*'
sudo apt install nvidia-driver-515

I am using nomodeset opts now...

[Richard (r-todd)]

I have this same error after unattended install upgraded from kernel 5.15.0-50 to 5.15.0.52 last night

- Ubuntu 22.04.1
- kernel 5.15.0-52-generic
- Nvidia Driver: 515.65.01
- Card: RTX 2070 SUPER

currently kernel 5.15.0-50 still boots

work around: nomodeset boot option

[Fab Ltm (frozenpeak)]

Same problem on my Dell Inspiron 16 laptop:

- Ubuntu 22.04.1
- Kernel 5.15.0-50-generic
- Nvidia Driver: 515.76
- Card: RTX 3060

The boot option "nomodeset" added in grub settings seems to solve temporaly the problem.

[vodopad27 (family-gan)]

I did fresh installation of Ubuntu 22.10. Issue still occur on fresh OS as well.

[jeremyszu (os369510)]

not sure if anyone reports here are same as #10 log.

if anyone got the nvidia stuck because of gpu-manager(#10).

Is it possible to share the /var/log/gpu-manager.log ?

not sure if the system is ssh-able when issue happens but if yes, it's easy to get the log.

or boot to multi-user.target (if it's reachable), then record the content of /var/log/gpu-manager.log when entering graphic.target.

or disable gpu-manager and start it after gnome-shell?

If we can know which command issued by gpu-manager then we may have some ideas.

[vodopad27 (family-gan)]

/var/log/gpu-manager.log

[vodopad27 (family-gan)]

Also i found the solution!

Listen:
1) I boot with latest kernel
2) After booting i do not see any output in monitor, just 'no signal'
3) Do not reboot! Go away for 2-3 minutes! Do not touch to PC
4) And after that all works! I see login screen and able to login! Just wait 2-3 min!

I attached journalctl -b

Someone, please, try this solution and confirm. Is this behavior reproduced on your system?

Not cool to wait 2-3 min, but works! There is some bug that should be fixed... Probably logs bellow can help to analyze problem?

[jeremyszu (os369510)]

looks like not u-d-c's issue because logind also got stuck when operates with nvidia.

I think either trying drm-tip / bisect kernel / reports it to nvidia may help.

[vodopad27 (family-gan)]

I created topic on Nvidia forum as well:
https://www.nvidia.com/en-us/geforce/forums/game-ready-drivers/13/503446/nvidia-drivers-not-work-on-ubuntu-22042210-mass-pr/

[vodopad27 (family-gan)]

https://forums.developer.nvidia.com/t/significant-delay-in-loading-xserver-with-proprietary-driver-on-kernel-version-above-5-15-0-48/ posted here as recommended on nvidia-forum.

[vodopad27 (family-gan)]

I found the solution!

On logs i saw this:
окт 31 19:45:11 vodka-PC kernel: i915 0000:00:02.0: [drm] *ERROR* Unclaimed access detected prior to suspending

And just for test i decided to disable internal videocard (HD 730) and problem has gone! Now i have fast boot and no issue! I rebooted two times and all fine.

I disabled internal videocard on motherboard bios. Please, try and confirm solution.

Probably there is no only nvidia issue, the issue related to intel drivers + nvidia drivers together?

UPD: also one person confirmed the solution:
https://forums.developer.nvidia.com/t/significant-delay-in-loading-xserver-with-proprietary-driver-on-kernel-version-above-5-15-0-48/232232/12?u=family-gan

[vodopad27 (family-gan)]

+ added xserver-xorg-video-intel to related packages. Probably related with issue. As far as i know xserver-xorg-video-intel package contain intel driver package.

[Mathieu letombe (mletombe)]

It seems OK for me. I was still starting my TV 10s after the PC and I didn't want to install any updates. I finally did it because there was nvidia-graphics-drivers-515 - 515.76+really.515.65.01-0ubuntu1 which was a fix for https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-515/+bug/1993041. This issue seems specific to RTX30xx cards plugged on HDMI, this is my case. I rebooted, it works even if I didn't shut my TV off.

[Joseph McKittrick (brivate)]

I have an AMD CPU and my Motherboard has no integrated graphics and I am still experiencing this problem.

Update-- Can confirm @mletombe that 520.56.06-0ubuntu0.22.04.1 solved my issues.

[Linuxium (linuxium.com.au)]

The latest Nvidia BETA driver (https://www.nvidia.com/Download/driverResults.aspx/194278/en-us/):

Linux X64 (AMD64/EM64T) Display Driver
Version: 525.53 BETA
Release Date: 2022.11.10
Operating System: Linux 64-bit
Language: English (US)
File Size: 394.19 MB

includes:

Fixed a regression in 515.76 that caused blank screens and hangs when starting an X server on RTX 30 series GPUs in some configurations where the boot display is connected via HDMI.

and this did indeed fix this issue for me.

Alternatively, keep an eye on https://launchpad.net/~graphics-drivers/+archive/ubuntu/ppa and when updated with version 525 do
sudo add-apt-repository ppa:graphics-drivers/ppa
sudo apt update
and then install the Ubuntu packaged driver for your series.

[Launchpad Janitor (janitor)]

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