Black screen, i915 not loaded with 5.15.0-41 (but 5.15.0-40 works)

[Leo Kotschenreuther]

After upgrading my Linux kernel to version 5.15.0-41, I am no longer able to get into the UI. After booting and unlocking the disk I end up on a black screen with a blinking underscore in the top left corner.

ProblemType: Bug
DistroRelease: Ubuntu 22.04
Package: gdm3 42.0-1ubuntu7
ProcVersionSignature: Ubuntu 5.15.0-41.44-generic 5.15.39
Uname: Linux 5.15.0-41-generic x86_64
NonfreeKernelModules: nvidia_modeset nvidia
ApportVersion: 2.20.11-0ubuntu82.1
Architecture: amd64
CasperMD5CheckResult: pass
Date: Thu Jul 21 12:14:04 2022
InstallationDate: Installed on 2021-12-03 (229 days ago)
InstallationMedia: Ubuntu 21.10 "Impish Indri" - Release amd64 (20211012)
ProcEnviron:
 TERM=linux
 SHELL=/usr/bin/zsh
 PATH=(custom, user)
 LANG=en_US.UTF-8
 XDG_RUNTIME_DIR=<set>
SourcePackage: gdm3
UpgradeStatus: No upgrade log present (probably fresh install)
mtime.conffile..etc.gdm3.custom.conf: 2022-07-21T11:41:09.164904
---
ProblemType: Bug
ApportVersion: 2.20.11-0ubuntu82.1
Architecture: amd64
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC1: leo 2796 F.... pulseaudio
 /dev/snd/controlC0: leo 2796 F.... pulseaudio
CRDA: N/A
CasperMD5CheckResult: pass
CurrentDesktop: ubuntu:GNOME
DistroRelease: Ubuntu 22.04
InstallationDate: Installed on 2021-12-03 (241 days ago)
InstallationMedia: Ubuntu 21.10 "Impish Indri" - Release amd64 (20211012)
MachineType: Dell Inc. XPS 15 9510
NonfreeKernelModules: nvidia_modeset nvidia
Package: linux (not installed)
ProcEnviron:
 TERM=xterm-256color
 PATH=(custom, no user)
 XDG_RUNTIME_DIR=<set>
 LANG=en_US.UTF-8
 SHELL=/usr/bin/zsh
ProcFB: 0 i915drmfb
ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-5.15.0-41-generic root=/dev/mapper/vgubuntu-root ro quiet splash vt.handoff=7
ProcVersionSignature: Ubuntu 5.15.0-41.44-generic 5.15.39
RelatedPackageVersions:
 linux-restricted-modules-5.15.0-41-generic N/A
 linux-backports-modules-5.15.0-41-generic N/A
 linux-firmware 20220329.git681281e4-0ubuntu3.3
Tags: jammy
Uname: Linux 5.15.0-41-generic x86_64
UpgradeStatus: No upgrade log present (probably fresh install)
UserGroups: adm cdrom dip docker libvirt lpadmin lxd plugdev sambashare sudo
_MarkForUpload: True
dmi.bios.date: 05/09/2022
dmi.bios.release: 1.10
dmi.bios.vendor: Dell Inc.
dmi.bios.version: 1.10.1
dmi.board.name: 0WH7HY
dmi.board.vendor: Dell Inc.
dmi.board.version: A00
dmi.chassis.type: 10
dmi.chassis.vendor: Dell Inc.
dmi.modalias: dmi:bvnDellInc.:bvr1.10.1:bd05/09/2022:br1.10:svnDellInc.:pnXPS159510:pvr:rvnDellInc.:rn0WH7HY:rvrA00:cvnDellInc.:ct10:cvr:sku0A61:
dmi.product.family: XPS
dmi.product.name: XPS 15 9510
dmi.product.sku: 0A61
dmi.sys.vendor: Dell Inc.

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

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 1982477

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.

[Leo Kotschenreuther (leokotschenreuther)]

I am unable to run apport-collect because the network is also unavailable with the new kernel version. Is there a way for me to generate the information without an internet connection (e.g. by using apport-cli) and to upload this information separately once I rebooted into the last working kernel version?

[Daniel van Vugt (vanvugt)]

Please:

1. Verify that booting the previous kernel 5.15.0-40 *does* work.

2. Verify that the system isn't just booting to the wrong VT. On the black screen press: Ctrl+Alt+F4, Ctrl+Alt+F1

3. Collect a log from when the bug last happened. If that's the current boot then run:

   journalctl -b0 > journal.txt

   If the bug last happened on the previous boot then please run:

   journalctl -b-1 > prevboot.txt

4. Also run:

   lspci -k > lspci.txt

5. Attach the resulting text files here.

[Daniel van Vugt (vanvugt)]

The problem might also be that the nvidia driver didn't get rebuilt for the new kernel. For whatever driver version you have installed, please run either:

  sudo apt install --reinstall nvidia-dkms-515

  sudo apt install --reinstall nvidia-dkms-510

[Leo Kotschenreuther (leokotschenreuther)]

The previous kernel 5.15.0-40 worked, at least in regards to having a working UI for me. It had other problems that I didn't get around to reporting yet (such as no network nor bluetooth being available). Should I report separate bugs for these?
I am currently using the kernel version 5.15.0-39 since that is the latest kernel version where everything works as expected.

When I press Ctrl+Alt+F1, nothing happens. When I press Ctrl+Alt+F4 I get to the terminal which I use to login and run all the commands.

I attached the journal and lspci files.

In terms of rebuilding the nvidia driver, that didn't make a difference. Neither nvidia-dkms-470 (which I apparently was using) nor switching to nvidia-dkms-515 and rebuilding that.

[Daniel van Vugt (vanvugt)]

The main error seems to be:

Jul 22 10:13:29 leo-home /usr/libexec/gdm-x-session[1713]: Fatal server error:
Jul 22 10:13:29 leo-home /usr/libexec/gdm-x-session[1713]: (EE) Cannot run in framebuffer mode. Please specify busIDs for all framebuffer devices

So I suspect there's an installation problem with the Nvidia driver. Please try using the 'Additional Drivers' app to reinstall it.

[Leo Kotschenreuther (leokotschenreuther)]

I already tried doing that using the 'Additional Drivers' app to no avail. One thing I also already attempted was to try out different Nvidia driver versions, but nothing changes.
* NVIDIA driver metapackage from nvidia-driver-515
* NVIDIA driver metapackage from nvidia-driver-510
* NVIDIA driver metapackage from nvidia-driver-470
* NVIDIA Server Driver metapackage from nvidia-driver-515-server

[Daniel van Vugt (vanvugt)]

Reviewing the log again, you have the nvidia-drm driver providing graphics device /dev/dri/card0. The failure only occurs in Xorg.

It looks like Xorg might be getting confused by a broken config file in /usr/share/X11/xorg.conf.d/ so please try:

1. Remove any Nvidia config files from /usr/share/X11/xorg.conf.d/
2. Reinstall the Nvidia driver.

[Leo Kotschenreuther (leokotschenreuther)]

That didn't help either.

I deleted all files with the word nvidia in the filename in:
/usr/share/X11/xorg.conf.d/
And then reinstalled the driver with
sudo apt install --reinstall nvidia-dkms-515

When I boot into the 5.15.0-41 kernel version, I still get the same behaviour as before.

[Daniel van Vugt (vanvugt)]

That's the latest Nvidia driver and the latest kernel, so I am wondering what else is unique about this system since no one else seems to be reporting the bug.

[Daniel van Vugt (vanvugt)]

Something strange in comment #7 - Your Intel GPU has no kernel driver loaded. And it's usually the only one wired to the screen. So that would explain a lot. It would also explain why I don't see other reports of the problem against the Nvidia driver if it's really an Intel GPU problem.

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

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 1982477

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.

[Leo Kotschenreuther (leokotschenreuther)]

My network driver also has problems with the new kernel version. I assume that running apport-collect will only provide useful information if I run it while running the problematic kernel version but that won't work because of the lack of a working network driver.
What commands shall I run to collect logs? I will then attach them here after rebooting into the older kernel version that works.

[Daniel van Vugt (vanvugt)]

Yes please boot a kernel that works (like 5.15.0-40) and then run:

  lspci -k > lspci-nobug.txt
  journalctl -b0 > journal-nobug.txt

and attach the resulting text files here.

[Leo Kotschenreuther (leokotschenreuther)]

Files attached

[Stefan Bader (smb)]

To double check: did linux-modules-extra get installed for the latest kernel version? (dpkg -l|grep linux-modules-extra)

[Daniel van Vugt (vanvugt)]

Yeah I can see i915 working properly in #17 and #18.

[Leo Kotschenreuther (leokotschenreuther)]

linux-modules-extra did not get installed for the latest kernel version.
The latest version I have it installed for is 5.15.0-39-generic which is the latest Kernel version where everything works for me (including ethernet, wifi and bluetooth).

[Leo Kotschenreuther (leokotschenreuther)]

I will give installing linux-modules-extra-5.15.0-41-generic a try.

[Daniel van Vugt (vanvugt)]

Like Stefan, I am also wondering about the statement:

> After upgrading my Linux kernel to version 5.15.0-41

This bug could occur if you manually installed the new kernel image package without the associated modules packages:

https://launchpad.net/ubuntu/+archive/primary/+files/linux-modules-5.15.0-41-generic_5.15.0-41.44_amd64.deb

https://launchpad.net/ubuntu/+archive/primary/+files/linux-modules-extra-5.15.0-41-generic_5.15.0-41.44_amd64.deb

[Leo Kotschenreuther (leokotschenreuther)]

Installing linux-modules-extra-5.15.0-41-generic resolved all problems. I now have a working UI, ethernet, wifi and bluetooth!

I followed the automated update process for installing the new kernel version, I don't remember having taken any manual steps.

Thanks so much for the help!

[Stefan Bader (smb)]

Normally this all should be driven by having linux-generic installed. That is a package which updates all packages related to the kernel at once. Minimally there should be linux-image-generic. But preferred is linux-generic since that also updates headers which are required for DKMS builds for example.

[Daniel van Vugt (vanvugt)]

OK so it was just "graphics driver not installed" :)

I'm not sure what went wrong in the upgrade process but if we care to find out then the answer might be in /var/log/apt/history* if you can find the log containing the original automatic update to 5.15.0-41

[Leo Kotschenreuther (leokotschenreuther)]

Attached is the history.log file, I think this is the one where the upgrade happened.

[Daniel van Vugt (vanvugt)]

Yes it appears your update on 2022-07-12 15:31:55 included most of the kernel 5.15.0-41 packages but omitted linux-modules-extra. So the Intel graphics driver was never installed.

Maybe the reason for not including linux-modules-extra was the presence of linux-modules-nvidia-470 ?

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

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 1982477

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.

[Steve Langasek (vorlon)]

Rather, this points to a problem of none of the kernel metapackages being installed except for the Nvidia hwe metapackage. So the upgrade only pulls in the new kernel ABI of those packages that are needed via dependency of the nvidia driver.

Why is the linux-generic package not installed here?

[Daniel van Vugt (vanvugt)]

The top of the page suggests a manual installation:

> After upgrading my Linux kernel to version 5.15.0-41

but comment #25 contradicts that.

[Daniel van Vugt (vanvugt)]

Being a dist upgrade from 21.10 might also be relevant.

[Leo Kotschenreuther (leokotschenreuther)]

apport information

[Leo Kotschenreuther (leokotschenreuther)]

apport information

[Leo Kotschenreuther (leokotschenreuther)]

apport information

[Leo Kotschenreuther (leokotschenreuther)]

apport information

[Leo Kotschenreuther (leokotschenreuther)]

apport information

[Leo Kotschenreuther (leokotschenreuther)]

apport information

[Leo Kotschenreuther (leokotschenreuther)]

apport information

[Leo Kotschenreuther (leokotschenreuther)]

apport information

[Leo Kotschenreuther (leokotschenreuther)]

apport information

[Leo Kotschenreuther (leokotschenreuther)]

apport information

[Leo Kotschenreuther (leokotschenreuther)]

apport information

[Leo Kotschenreuther (leokotschenreuther)]

apport information

[Leo Kotschenreuther (leokotschenreuther)]

apport information

[Leo Kotschenreuther (leokotschenreuther)]

apport information

[Leo Kotschenreuther (leokotschenreuther)]

apport information

[Leo Kotschenreuther (leokotschenreuther)]

apport information

[Leo Kotschenreuther (leokotschenreuther)]

apport information

[Leo Kotschenreuther (leokotschenreuther)]

apport information

[Leo Kotschenreuther (leokotschenreuther)]

As I stated earlier, I used the Software Updater which pops up automatically for me whenever there is new software to install/ upgrade. Sorry for the confusion.

[Steve Langasek (vorlon)]

Leo, I've looked at the apt logs you've provided, and they explain why things were broken for you after upgrade, but the key question remains unanswered:

Why is the linux-generic package not installed here?

Nothing in your history.log, starting July 5, shows it being removed. It (or an equivalent metapackage) is *always* installed as part of an install. So there doesn't appear to be a kernel bug here, and if there is a bug in Ubuntu (as opposed to user error by forcing the removal of this metapackage), we need to see when the package disappeared from your system in order to further analyze it.

[Daniel van Vugt (vanvugt)]

Does linux-generic-hwe-20.04 automatically upgrade to linux-generic?

[Steve Langasek (vorlon)]

On Wed, Aug 03, 2022 at 03:00:41AM -0000, Daniel van Vugt wrote:
> Does linux-generic-hwe-20.04 automatically upgrade to linux-generic?

$ apt-cache show linux-generic-hwe-20.04 | grep Depends
Depends: linux-generic-hwe-22.04
Depends: linux-generic-hwe-22.04
$

I don't know offhand why that particular migration has been chosen instead
of linux-generic. Regardless, if an appropriate metapackage were present,
it would have shown up in history.log.

[Leo Kotschenreuther (leokotschenreuther)]

I don't really know what else to add to the matter.
Should we close this bug for now?
I can always reopen it should the issue happen again on the next automatic Kernel update. At least this time I know which package to look for should there be any issues.

[Daniel van Vugt (vanvugt)]

If you can find older versions of /var/log/apt/history* that contain the string "linux-generic" then that might help us understand what happened.

[Leo Kotschenreuther (leokotschenreuther)]

Unfortunately I accidentally deleted the history log file from June. I think this would have been the one where we could have found more information.

[Daniel van Vugt (vanvugt)]

Then we will never know... unless someone can reproduce the bug (from a dist upgrade presumably).

If that doesn't happen in the next 60 days (and there are no further comments) then the bug will close automatically.

[Launchpad Janitor (janitor)]

[Expired for linux (Ubuntu) because there has been no activity for 60 days.]