Ubuntu 22.04.1 LTS does not boot after update 5.15.0-48 -> 5.15.0-56 - old kernel boots fine

[Michael Soegtrop]

My Ubuntu 22.04.1 LTS auto updated from kernel 5.15.0-48 to 5.15.0-56.

The new kernel does not boot. It stops right after it says the disk is clean. The cursor stops blinking less than 1 second after that and there is no additional message.

The old kernel boots fine. The "ubuntu-bug" report was done with the old kernel.

I am new to Ubuntu bug reporting and hope the information created by `ubuntu-bug` is added, even though I can't see it here.

ProblemType: Bug
DistroRelease: Ubuntu 22.04
Package: linux-image-5.15.0-48-generic 5.15.0-48.54
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: 2.20.11-0ubuntu82.2
Architecture: amd64
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC0: michael 2477 F.... pulseaudio
CRDA: N/A
CasperMD5CheckResult: pass
CurrentDesktop: ubuntu:GNOME
Date: Sun Dec 4 11:13:53 2022
InstallationDate: Installed on 2022-09-19 (75 days ago)
InstallationMedia: Ubuntu 22.04.1 LTS "Jammy Jellyfish" - Release amd64 (20220809.1)
MachineType: Supermicro M12SWA-TF
ProcFB:
 0 EFI VGA
 1 astdrmfb
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-5.15.0-48-generic root=UUID=33a91d45-6bee-4e70-bc21-7956db527e81 ro quiet splash vt.handoff=7
RelatedPackageVersions:
 linux-restricted-modules-5.15.0-48-generic N/A
 linux-backports-modules-5.15.0-48-generic N/A
 linux-firmware 20220329.git681281e4-0ubuntu3.7
RfKill:

SourcePackage: linux
UpgradeStatus: No upgrade log present (probably fresh install)
dmi.bios.date: 03/18/2022
dmi.bios.release: 2.10
dmi.bios.vendor: American Megatrends International, LLC.
dmi.bios.version: 2.0a
dmi.board.asset.tag: Default string
dmi.board.name: M12SWA-TF
dmi.board.vendor: Supermicro
dmi.board.version: 1.01
dmi.chassis.asset.tag: Default string
dmi.chassis.type: 17
dmi.chassis.vendor: Supermicro
dmi.chassis.version: 0123456789
dmi.modalias: dmi:bvnAmericanMegatrendsInternational,LLC.:bvr2.0a:bd03/18/2022:br2.10:svnSupermicro:pnM12SWA-TF:pvr123456789:rvnSupermicro:rnM12SWA-TF:rvr1.01:cvnSupermicro:ct17:cvr0123456789:sku091715D9:
dmi.product.family: SMC M12
dmi.product.name: M12SWA-TF
dmi.product.sku: 091715D9
dmi.product.version: 123456789
dmi.sys.vendor: Supermicro

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

This change was made by a bot.

[Michael Soegtrop (msoegtrop)]

I added a boot log for the new kernel created with "journalctl -b -1".

[Kai-Heng Feng (kaihengfeng)]

Dez 04 10:35:15 TRUB22 gdm3[1980]: modprobe: FATAL: Module nvidia not found in directory /lib/modules/5.15.0-56-generic

[Michael Soegtrop (msoegtrop)]

So I guess the nVidia driver was not compiled for the new kernel. Is this something which should happen automatically during an update? I am using Ubuntu with the proprietary nVidia driver since 14.X or so and never had such an issue.

Is this an issue I should report at nVidia?

What is the recommended procedure to fix this?

[Alberto Milone (albertomilone)]

Michael: how did you install the NVIDIA driver?

[Michael Soegtrop (msoegtrop)]

I installed it via Ubuntu's "use proprietary drivers" function. I double checked that I have no trace of a downloaded nVidia driver installer on my machine.

[Michael Soegtrop (msoegtrop)]

One more note: my motherboard has a BMC which provides a pseudo VGA card I can access e.g. via the web interface of the BMC. To Linux this is a secondary display card and screen.

[Michael Soegtrop (msoegtrop)]

A few more notes (with a manual fix):

After an update to kernel 5.15.0-58 I again had the same issue.

The (slightly shortened) state the Ubuntu updater left my system in after the kernel update is:
```
dpkg-query --list '*' | grep kernel
ii linux-headers-5.15.0-48 5.15.0-48.54 all Header files related to Linux kernel version 5.15.0
ii linux-headers-5.15.0-48-generic 5.15.0-48.54 amd64 Linux kernel headers for version 5.15.0 on 64 bit x86 SMP
ii linux-headers-5.15.0-58 5.15.0-58.64 all Header files related to Linux kernel version 5.15.0
ii linux-headers-5.15.0-58-generic 5.15.0-58.64 amd64 Linux kernel headers for version 5.15.0 on 64 bit x86 SMP
ii linux-headers-generic-hwe-22.04 5.15.0.58.56 amd64 Generic Linux kernel headers
ii linux-image-5.15.0-48-generic 5.15.0-48.54 amd64 Signed kernel image generic
ii linux-image-5.15.0-58-generic 5.15.0-58.64 amd64 Signed kernel image generic
ii linux-image-generic-hwe-22.04 5.15.0.58.56 amd64 Generic Linux kernel image
ii linux-modules-5.15.0-48-generic 5.15.0-48.54 amd64 Linux kernel extra modules for version 5.15.0 on 64 bit x86 SMP
ii linux-modules-5.15.0-58-generic 5.15.0-58.64 amd64 Linux kernel extra modules for version 5.15.0 on 64 bit x86 SMP
ii linux-modules-extra-5.15.0-48-generic 5.15.0-48.54 amd64 Linux kernel extra modules for version 5.15.0 on 64 bit x86 SMP
ii linux-modules-extra-5.15.0-58-generic 5.15.0-58.64 amd64 Linux kernel extra modules for version 5.15.0 on 64 bit x86 SMP
ii linux-modules-nvidia-515-5.15.0-48-generic 5.15.0-48.54 amd64 Linux kernel nvidia modules for version 5.15.0-48
ii linux-objects-nvidia-515-5.15.0-48-generic 5.15.0-48.54 amd64 Linux kernel nvidia modules for version 5.15.0-48 (objects)
ii linux-signatures-nvidia-5.15.0-48-generic 5.15.0-48.54 amd64 Linux kernel signatures for nvidia modules for version 5.15.0-48-generic
```
So the nvidia modules matching the kernel are simply missing.

Manually installing them (with the old 48 kernel booted) with
```
sudo apt install linux-signatures-nvidia-5.15.0-58-generic
sudo apt install linux-modules-nvidia-515-5.15.0-58-generic
sudo apt install linux-objects-nvidia-515-5.15.0-58-generic (nothing extra installed)
```
fixed the boot issue with kernel 58.

A possible toot cause:

In the old days dkms was supposed to handle this. Apparently in recent Ubuntu the Ubuntu supplied nvidia drivers are installed as binary packages. On my system dkms is not installed. Manually installing dkms and running "dmks status" brings no result. This decision of Ubuntu is menti...

[Michael Soegtrop (msoegtrop)]

Apparently this has been fixed in `5.19.0-32-generic`. The automatic update to this kernel version did install the proper nvidia drivers without manual interaction.

I am not 100% sure, though, if this is possibly an effect of installing the missing modules manually in the previous version. Possibly these then just get updated on their own right rather than as some dependency or so. I don't know much about how this works exactly.