Nvidia-modeset/kthread_q 100% CPU + frozen computer (nvidia proprietary driver)

Hi,

Since I upgraded to fedora 44, I experience computer freeze under
Plasma (wayland). Sometimes when in screen lock mode, I cannot log
back again because the computer is not responding or it happens just
after I unlock the screen. I can still log in via ssh and see that
nvidia-modeset/kthread_q is using 100% of the CPU.

 PID USER      PR  NI    VIRT    RES    SHR S  %CPU  %MEM     TIME+ COMMAND
1073 root      20   0       0      0      0 R  98,8   0,0  17:17.71
[nvidia-modeset/kthread_q]

The process cannot be killed with signal 9 or 15 and the the computer
does not even turn off with shutdown command. The only thing I can do
is to run the sysreq commands to reboot with
Alt+Sysreq+REISUB.

I use the proprietary nvidia drivers from rpmfusion (610.43.03). My
laptop runs an AMD CPU (Ryzen AI 7 350 w/ Radeon 860M) with an NVIDIA
GPU (GeForce RTX 5060). I recently upgraded the BIOS of my computer to
the latest available but it did not improve.

Last time it happened, dmesg|grep error gave:

ACPI Error: Aborting method \\\_SB.ACDC.RTWT due to previous error (AE_NOT_EXIST) (20251212/psparse-529)
ACPI Error: Aborting method \\\_SB.ACDC.\_GRT due to previous error (AE_NOT_EXIST) (20251212/psparse-529)
ACPI Error: Aborting method \\\_SB.WMID.WQBZ due to previous error (AE_AML_BUFFER_LIMIT) (20251212/psparse-529)
ACPI Error: Aborting method \\\_SB.WMID.WQBE due to previous error (AE_AML_BUFFER_LIMIT) (20251212/psparse-529)
hp_bioscfg: Returned error 0x3, "Invalid command value/Feature not supported"

It is possible that those are the latest logs in journalctl just
before the freeze:

kernel: nvidia 0000:05:00.0: Enabling HDA controller 
audit: BPF prog-id=114 op=LOAD 
systemd\[1\]: Starting fprintd.service - Fingerprint Authentication Daemon... 
systemd\[1\]: Started fprintd.service - Fingerprint Authentication Daemon. 
audit\[1\]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=fprintd comm="syste
md" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' 
alpha systemd\[1\]: Started dbus-:1.3-org.kde.powerdevil.backlighthelper@23.service. 
audit\[1\]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=dbus-:1.3-org.kde.p
owerdevil.backlighthelper@23 comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' 
audit\[68171\]: AUDIT1100 pid=68171 uid=1000 auid=1000 ses=3 subj=unconfined_u:unconfined_r:chkpwd_t:s0-s0:c0.c1023 msg='op=PAM:uni
x_chkpwd acct="fred" exe="/usr/bin/unix_chkpwd" hostname=? addr=? terminal=? res=success' 
alpha systemd\[1\]: dbus-:1.3-org.kde.powerdevil.backlighthelper@23.service: Deactivated successfully. 
audit\[1\]: SERVICE_STOP pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:init_t:s0 msg='unit=dbus-:1.3-org.kde.po
werdevil.backlighthelper@23 comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' 
systemd\[2194\]: app-org.kde.dolphin@5bc932b56eb94c20b1cf116060b04205.service: Consumed 10.477s CPU time over 4h 24min 14.502s wall
 clock time, 62.5M memory peak, 26.1M memory swap peak. 
kwin_wayland\[2553\]: Libinput: event6  - SYNA3501:00 06CB:CFC5 Touchpad: kernel bug: Touch jump detected and discarded. 
                                          See https://wayland.freedesktop.org/libinput/doc/1.31.3/touchpad-jumping-cursors.html for details

Also here are my boot options:

$ cat /proc/cmdline
BOOT_IMAGE=(hd0,gpt1)/vmlinuz-7.1.5-201.fc44.x86_64 root=UUID=31d7af0b-b478-4948-8c95-da71df62238c ro rootflags=subvol=root resume=UUID=797f8139-80b9-4f52-8e52-07a2127e90d1 rhgb quiet kvm.enable_virt_at_load=0 rd.driver.blacklist=nouveau,nova_core modprobe.blacklist=nouveau,nova_core

Anyone experiencing the same issue?
Any advice on what to look at when this happens?

Thanks,

F

I do not have a nvidia gpu and can only offer suggestions.
Nvidia update the 610 driver with a number of bug fixes.
Maybe when rpmfusion has that available it might help.
The version released by nvidia is 610.57.04 according to phronix.com.