htop shows 100% usage of systemd-udevd, constant spamming of nvidia in journalctl

[Lyubomir]

I'm attaching a perf profile.

ProblemType: Bug
DistroRelease: Ubuntu 20.04
Package: systemd 245.4-4ubuntu3.3
ProcVersionSignature: Ubuntu 5.4.0-54.60-lowlatency 5.4.65
Uname: Linux 5.4.0-54-lowlatency x86_64
ApportVersion: 2.20.11-0ubuntu27.12
Architecture: amd64
CasperMD5CheckResult: skip
Date: Sun Nov 22 13:57:47 2020
InstallationDate: Installed on 2020-02-08 (287 days ago)
InstallationMedia: Ubuntu 19.10 "Eoan Ermine" - Release amd64 (20191017)
MachineType: Acer Aspire A515-51G
ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-5.4.0-54-lowlatency root=/dev/mapper/vgubuntu-root ro quiet splash resume=UUID=f0e63db4-d3af-4977-a82a-5a46a3c677ae lockdown=confidentiality intel_iommu=on iommu=pt vfio-pci.ids=10de:1d10 vt.handoff=7
SourcePackage: systemd
SystemdDelta:
 [EXTENDED] /usr/lib/systemd/system/dnscrypt-proxy.service → /etc/systemd/system/dnscrypt-proxy.service.d/override.conf
 [EXTENDED] /usr/lib/systemd/system/rc-local.service → /usr/lib/systemd/system/rc-local.service.d/debian.conf
 [EXTENDED] /usr/lib/systemd/system/user@.service → /usr/lib/systemd/system/user@.service.d/timeout.conf

 3 overridden configuration files found.
UpgradeStatus: Upgraded to focal on 2020-10-23 (29 days ago)
dmi.bios.date: 01/03/2019
dmi.bios.vendor: Insyde Corp.
dmi.bios.version: V2.02
dmi.board.asset.tag: Type2 - Board Serial Number
dmi.board.name: Charmander_KL
dmi.board.vendor: KBL
dmi.board.version: V2.02
dmi.chassis.type: 10
dmi.chassis.vendor: Acer
dmi.chassis.version: V2.02
dmi.modalias: dmi:bvnInsydeCorp.:bvrV2.02:bd01/03/2019:svnAcer:pnAspireA515-51G:pvrV2.02:rvnKBL:rnCharmander_KL:rvrV2.02:cvnAcer:ct10:cvrV2.02:
dmi.product.family: Aspire 5
dmi.product.name: Aspire A515-51G
dmi.product.sku: 0000000000000000
dmi.product.version: V2.02
dmi.sys.vendor: Acer
mtime.conffile..etc.apport.crashdb.conf: 2020-05-15T09:41:35.683742
mtime.conffile..etc.systemd.sleep.conf: 2020-09-26T18:33:31.018002

[Lyubomir (mystiquewolf)]

Today it again does this, but this time immediately after i SIGKILL the process it again resumes with 100% CPU.

[Dan Streetman (ddstreet)]

Your problem appears to be with the nvidia driver and/or service(s), not with systemd; do you have any 3rd-party nvidia packages or tools installed, or are you using the official ubuntu-provided nvidia packages?

[Lyubomir (mystiquewolf)]

I have not installed any unofficial ppa or manual drivers. The only ones that are supposed to be installed are the official binary nvidia drivers. However, i remember that recently through "Software & Updates -> Additional drivers" i clicked a newer version (455, IIRC), it showed a progress bar, but i decided that i'll take many MBs and i' currently on metered connection, so i clicked "Revert" or "Cancel" or something like that. Then i clicked "Close".

I only have these packages installed:
libnvidia-cfg1-450/focal-updates,now 450.80.02-0ubuntu0.20.04.2 amd64 [installed]
libnvidia-common-450/focal-updates,focal-updates,now 450.80.02-0ubuntu0.20.04.2 all [installed]
libnvidia-compute-450/focal-updates,now 450.80.02-0ubuntu0.20.04.2 amd64 [installed]
libnvidia-compute-450/focal-updates,now 450.80.02-0ubuntu0.20.04.2 i386 [installed,automatic]
libnvidia-decode-450/focal-updates,now 450.80.02-0ubuntu0.20.04.2 amd64 [installed]
libnvidia-decode-450/focal-updates,now 450.80.02-0ubuntu0.20.04.2 i386 [installed]
libnvidia-encode-450/focal-updates,now 450.80.02-0ubuntu0.20.04.2 amd64 [installed]
libnvidia-encode-450/focal-updates,now 450.80.02-0ubuntu0.20.04.2 i386 [installed]
libnvidia-extra-450/focal-updates,now 450.80.02-0ubuntu0.20.04.2 amd64 [installed]
libnvidia-fbc1-450/focal-updates,now 450.80.02-0ubuntu0.20.04.2 amd64 [installed]
libnvidia-fbc1-450/focal-updates,now 450.80.02-0ubuntu0.20.04.2 i386 [installed]
libnvidia-gl-450/focal-updates,now 450.80.02-0ubuntu0.20.04.2 amd64 [installed]
libnvidia-gl-450/focal-updates,now 450.80.02-0ubuntu0.20.04.2 i386 [installed]
libnvidia-ifr1-450/focal-updates,now 450.80.02-0ubuntu0.20.04.2 amd64 [installed]
libnvidia-ifr1-450/focal-updates,now 450.80.02-0ubuntu0.20.04.2 i386 [installed]
nvidia-compute-utils-450/focal-updates,now 450.80.02-0ubuntu0.20.04.2 amd64 [installed]
nvidia-dkms-450/focal-updates,now 450.80.02-0ubuntu0.20.04.2 amd64 [installed]
nvidia-kernel-common-450/focal-updates,now 450.80.02-0ubuntu0.20.04.2 amd64 [installed]
nvidia-kernel-source-450/focal-updates,now 450.80.02-0ubuntu0.20.04.2 amd64 [installed]
nvidia-prime/focal,focal,now 0.8.14 all [installed]
nvidia-settings/focal-updates,now 440.82-0ubuntu0.20.04.1 amd64 [installed]
nvidia-utils-450/focal-updates,now 450.80.02-0ubuntu0.20.04.2 amd64 [installed]
xserver-xorg-video-nvidia-450/focal-updates,now 450.80.02-0ubuntu0.20.04.2 amd64 [installed]

[Dan Streetman (ddstreet)]

If you stopped mid-upgrade, I'd suggest finishing the package upgrade and reboot. If you still have problems with fully/correctly upgraded packages you will need to provide more logs; the journal for systemd-udevd is a good place to start.

[Lyubomir (mystiquewolf)]

Okay, so i installed manually through terminal nvidia-drivers-455, which removed the old 450 and installed itself, then i rebooted. After reboot now htop doesn't show 100% constant of systemd-udevd, but shows between 20% to 80%, the medium looks like 50% CPU usage. In Journalctl there is constant nagging:
ное 23 23:48:22 lyubomir-Aspire-A515-51G kernel: nvidia-nvlink: Nvlink Core is being initialized, major device number 236
ное 23 23:48:22 lyubomir-Aspire-A515-51G kernel: NVRM: The NVIDIA probe routine was not called for 1 device(s).
ное 23 23:48:22 lyubomir-Aspire-A515-51G kernel: NVRM: This can occur when a driver such as:
                                                    NVRM: nouveau, rivafb, nvidiafb or rivatv
                                                    NVRM: was loaded and obtained ownership of the NVIDIA device(s).
ное 23 23:48:22 lyubomir-Aspire-A515-51G kernel: NVRM: Try unloading the conflicting kernel module (and/or
                                                    NVRM: reconfigure your kernel without the conflicting
                                                    NVRM: driver(s)), then try loading the NVIDIA kernel module
                                                    NVRM: again.
ное 23 23:48:22 lyubomir-Aspire-A515-51G kernel: NVRM: No NVIDIA devices probed.
ное 23 23:48:22 lyubomir-Aspire-A515-51G kernel: nvidia-nvlink: Unregistered the Nvlink Core, major device number 236
ное 23 23:48:22 lyubomir-Aspire-A515-51G systemd-udevd[709]: nvidia: Process '/sbin/modprobe nvidia-uvm' failed with exit code 1.
ное 23 23:48:22 lyubomir-Aspire-A515-51G systemd[1]: nvidia-persistenced.service: Start request repeated too quickly.
ное 23 23:48:22 lyubomir-Aspire-A515-51G systemd[1]: nvidia-persistenced.service: Failed with result 'exit-code'.
ное 23 23:48:22 lyubomir-Aspire-A515-51G systemd[1]: Failed to start NVIDIA Persistence Daemon.
ное 23 23:48:22 lyubomir-Aspire-A515-51G boinc[2398]: No protocol specified
ное 23 23:48:22 lyubomir-Aspire-A515-51G boinc[2398]: No protocol specified
ное 23 23:48:22 lyubomir-Aspire-A515-51G kernel: nvidia-nvlink: Nvlink Core is being initialized, major device number 236
ное 23 23:48:22 lyubomir-Aspire-A515-51G kernel: NVRM: The NVIDIA probe routine was not called for 1 device(s).
ное 23 23:48:22 lyubomir-Aspire-A515-51G kernel: NVRM: This can occur when a driver such as:
                                                    NVRM: nouveau, rivafb, nvidiafb or rivatv
                                                    NVRM: was loaded and obtained ownership of the NVIDIA device(s).
ное 23 23:48:22 lyubomir-Aspire-A515-51G kernel: NVRM: Try unloading the conflicting kernel module (and/or
                                                    NVRM: reconfigure your kernel without the conflicting
                                                    NVRM: driver(s)), then try loading the NVIDIA kernel module
                                                    NVRM: again.
ное 23 23:48:22 lyubomir-Aspire-A515-51G kernel: NVRM: No NVIDIA devices probed.
ное 23 23:48:22 lyubomir-Aspire-A515-51G kernel: nvidia-nvlink: Unregistered the Nvlink Core, major device number 236
ное 23 23:48:22 lyubomir-...

[Lyubomir (mystiquewolf)]

Might be this bug: https://bugs.launchpad.net/ubuntu/+source/systemd/+bug/1655584
I use KVM/QEMU and load vfio-pci driver in place of NVidia driver.

[Lyubomir (mystiquewolf)]

Is anyone triaging these bugs, like, if you can't fix all of them at least acknowledge them. I'm not talking only for nvidia package bugs but for many others also. Somehow the Firefox bugzilla triage works better IMO.

[Lyubomir (mystiquewolf)]

I should also note that attaching my Android phone isn't detected/showed in Nautilus, but not sure if this has anything to do with systemd-udevd.