Public bug reported:
Summary
On a hybrid-graphics laptop with an RTX 5060 Laptop GPU, boot and shutdown hang
most of the time (very roughly 90% of attempts) on kernels 7.0.0-31 and later.
Boot stalls on "gpu-manager.service/start running" with plymouth-quit waiting
behind it. 7.0.0-30 did not have this problem.
Hardware / software
- Laptop: Lenovo Legion 5 16IAX10
- CPU: Intel Core i9 13th gen (HX series) with integrated graphics
- dGPU: NVIDIA GeForce RTX 5060 Laptop GPU (PCI 02:00.0), GPU mode set to Hybrid
- OS: Ubuntu 26.04, dual-boot with Windows
- Driver: nvidia-driver-595-open 595.91.07-0ubuntu0.26.04.1, with
linux-modules-nvidia-595-open-<kernel> prebuilt modules (no DKMS)
- Secure Boot: disabled
- Affected kernels: 7.0.0-31, 7.0.0-38
- Last known good kernel: 7.0.0-30
Symptoms
- Boot hangs at gpu-manager.service start; plymouth-quit waits on it
- Shutdown/restart also hangs most of the time
- Boot succeeds only occasionally, which suggests a race
Evidence from the journal of a hung boot (journalctl -b -1)
The nvidia, nvidia-modeset and nvidia-drm modules begin loading, and
gpu-manager.service starts at the same moment. After that, udev reports:
nvidia: Spawned process '/sbin/modprobe nvidia-drm' [1169] is taking longer
than 59s to complete.
systemd-udevd: nvidia: Worker [679] processing SEQNUM=5707 is taking a long
time.
No hung-task ("blocked for more than 120 seconds") trace was captured
because the system was restarted before that threshold. I can capture
one if needed.
Workaround being tested
Setting nvidia-drm.fbdev=0 (via /etc/modprobe.d, with the initramfs rebuilt and
the value confirmed as N in /sys/module/nvidia_drm/parameters/fbdev) has so far
produced no hangs across roughly 15 minutes of repeated boots and restarts.
This is not yet confirmed as a fix, and the root cause is unknown. An earlier
test with a mistyped kernel parameter had no effect (fbdev stayed Y), and boots
succeeded anyway, so the failure rate may be lower than initially estimated or
variable. A longer test is in progress.
Request
Please investigate whether a deadlock or race in nvidia-drm initialization
(possibly involving fbdev setup alongside the Intel GPU driver) was introduced
between kernel 7.0.0-30 and 7.0.0-31. I can provide further logs on request.
ProblemType: Bug
DistroRelease: Ubuntu 26.04
Package: nvidia-driver-595-open 595.91.07-0ubuntu0.26.04.1
ProcVersionSignature: Ubuntu 7.0.0-38.38-generic 7.0.14
Uname: Linux 7.0.0-38-generic x86_64
ApportVersion: 2.34.1-0ubuntu0.1
Architecture: amd64
CasperMD5CheckResult: pass
CurrentDesktop: ubuntu:GNOME
Date: Wed Oct 7 01:03:39 2026
InstallationDate: Installed on 2026-08-22 (45 days ago)
InstallationMedia: Ubuntu 26.04 "Resolute Raccoon" - Release amd64 (20260423.1)
ProcEnviron:
LANG=en_US.UTF-8
PATH=(custom, no user)
SHELL=/bin/bash
TERM=xterm-256color
XDG_RUNTIME_DIR=<set>
SourcePackage: nvidia-graphics-drivers-595
UpgradeStatus: No upgrade log present (probably fresh install)
** Affects: nvidia-graphics-drivers-595 (Ubuntu)
Importance: Undecided
Status: New
** Tags: amd64 apport-bug resolute wayland-session
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2169798
Title:
nvidia-drm modprobe hangs during boot/shutdown on hybrid laptop (RTX
5060, nvidia-driver-595-open), regression starting with kernel
7.0.0-31
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers-595/+bug/2169798/+subscriptions
--
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
