[Open] Nvidia reports NVKMS memory for GEM object failure; no bus fail

Yumasai · 2025-12-13 13:21:46

DISCLAIMER

I am new to the forums, so please bear with me whilst I try to report the problem and give information.

===============================

CONTEXT

Since december 1st 2025 I am experiencing issues regarding playing Total War Warhammer 3 and then alt-tabbing; it causes the game itself to freeze, and other applications freeze graphically as well. What I thus mean is that the process itself still runs, i.e. button presses still give audible feedback. The graphical output is just frozen. Nothing seems to fix the issue besides rebooting sddm, which obviously kills my current session. Game is not on unlimited vram

I have dug into journalctl and grep'd on nvidia. I noticed a particuliar entry on today's date, once the issue popped up again:

Dec 13 13:27:28 archlinux kernel: Command line: initrd=\amd-ucode.img initrd=\initramfs-linux.img root=PARTUUID=c6c23ccc-4121-4973-b01e-f3c16565972c zswap.enabled=0 rw rootfstype=ext4 nvidia_drm.modeset=1
Dec 13 13:27:28 archlinux kernel: Kernel command line: initrd=\amd-ucode.img initrd=\initramfs-linux.img root=PARTUUID=c6c23ccc-4121-4973-b01e-f3c16565972c zswap.enabled=0 rw rootfstype=ext4 nvidia_drm.modeset=1
Dec 13 13:27:30 archlinux kernel: nvidia: loading out-of-tree module taints kernel.
Dec 13 13:27:30 archlinux kernel: nvidia: module verification failed: signature and/or required key missing - tainting kernel
Dec 13 13:27:30 archlinux kernel: nvidia-nvlink: Nvlink Core is being initialized, major device number 239
Dec 13 13:27:30 archlinux kernel: nvidia 0000:2b:00.0: enabling device (0002 -> 0003)
Dec 13 13:27:30 archlinux kernel: nvidia 0000:2b:00.0: vgaarb: VGA decodes changed: olddecodes=io+mem,decodes=none:owns=none
Dec 13 13:27:31 archlinux kernel: nvidia-gpu 0000:2b:00.3: enabling device (0000 -> 0002)
Dec 13 13:27:31 archlinux systemd-modules-load[381]: Inserted module 'nvidia_uvm'
Dec 13 13:27:31 archlinux kernel: nvidia-modeset: Loading NVIDIA UNIX Open Kernel Mode Setting Driver for x86_64  580.105.08  Release Build  (root@)  
Dec 13 13:27:31 archlinux kernel: [drm] [nvidia-drm] [GPU ID 0x00002b00] Loading driver
Dec 13 13:27:32 archlinux kernel: nvidia-gpu 0000:2b:00.3: i2c timeout error e0000000
Dec 13 13:27:32 archlinux kernel: [drm] Initialized nvidia-drm 0.0.0 for 0000:2b:00.0 on minor 1
Dec 13 13:27:32 archlinux kernel: nvidia 0000:2b:00.0: vgaarb: deactivate vga console
Dec 13 13:27:32 archlinux kernel: fbcon: nvidia-drmdrmfb (fb0) is primary device
Dec 13 13:27:32 archlinux kernel: nvidia 0000:2b:00.0: [drm] fb0: nvidia-drmdrmfb frame buffer device
Dec 13 13:40:58 archlinux kernel: nvidia 0000:2b:00.0: Using 47-bit DMA addresses
Dec 13 13:41:06 archlinux kernel: [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00002b00] Failed to allocate NVKMS memory for GEM object
Dec 13 13:41:20 archlinux kernel: [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00002b00] Failed to allocate NVKMS memory for GEM object
Dec 13 13:41:25 archlinux kernel: [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00002b00] Failed to allocate NVKMS memory for GEM object
Dec 13 13:41:32 archlinux kernel: [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00002b00] Failed to allocate NVKMS memory for GEM object
Dec 13 13:41:32 archlinux kernel: [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00002b00] Failed to allocate NVKMS memory for GEM object
Dec 13 13:41:32 archlinux kernel: [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00002b00] Failed to allocate NVKMS memory for GEM object
Dec 13 13:41:32 archlinux kernel: [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00002b00] Failed to allocate NVKMS memory for GEM object
Dec 13 13:41:32 archlinux kernel: [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00002b00] Failed to allocate NVKMS memory for GEM object

===============================

SPECS

specs

Sat Dec 13 14:37:52 2025       
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 580.105.08             Driver Version: 580.105.08     CUDA Version: 13.0     |
+-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA GeForce RTX 2070 ...    Off |   00000000:2B:00.0  On |                  N/A |
|  0%   31C    P8              6W /  215W |     683MiB /   8192MiB |      1%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+

Last edited by Yumasai (2025-12-13 13:38:25)

seth · 2025-12-13 17:51:10

grep'd on nvidia

Don't. nvidia files the most important stuff under NVRM.
Are the timestamps there even relevant to the crash?

Also never post pictures of text, post the text.
The picture does also not say whether it's wayland or X11 - I'll assume wayland?

playing Total War Warhammer 3 and then alt-tabbing

Is it *only* "Total War Warhammer 3"? "Total War Warhammer 2 - Judgement Day" or other games are not affected?
Any of those https://aur.archlinux.org/packages?O=0&K=unigine ?

I guees "Total War Warhammer 3" implies steam and thus xwayland?
Does it also happen
1. on an X11 plasma session?
2. on a non-plasma wayland session? (weston will do but you can also try hyprland, labwc or sway for a more realistic test)

Yumasai · 2025-12-13 20:03:03

Hi Seth,

The timestamps are relevant, yes. 13:27:x is system start. 13:41:x is the issue arising.

Wayland, yes.

Well TW W3 has the issue beginning but obviously after that, discord, browser, etc all graphically freeze. The only thing still working is my bottom taskbar where I can right click on application icons.

Also, yes, I play it through Steam with Proton Experimental (so I assume xwayland).

I have not encountered it on X11 or non-plasma wayland. I have been considering Sway but I am a bit hesitant to do a full swap to it...

seth · 2025-12-13 21:15:50

but I am a bit hesitant to do a full swap to it...

This is not about selling you a new DE but to figure whether TWW3 has a problem w/ xwayland or kwin_wayland has a problem w/ TWW3
You shall just start a different wayland compositor, run the game and see whether you can make it crash.

Could be related to some kwin unredirection (plus the latest update of kwin_wayland seems to have problems anyway - you could attempt to downgrade it, but first
https://gitlab.archlinux.org/archlinux/ … mmits/main
Which one is it? Yesterday there were two updates addressing freezes specifically.

Yumasai · 2025-12-13 21:41:45

Hi Seth,

I have shifted to x11 for now, and playing WH3 seems to not make the issue reproducable; that said, the version I have is:

[yumasai@archlinux ~]$ kwin_x11 --version
kwin 6.5.4
QThreadStorage: entry 1 destroyed before end of thread 0x55875c89d810
QThreadStorage: entry 0 destroyed before end of thread 0x55875c89d810

seth · 2025-12-13 21:44:22

pacman -Qi kwin

The build matters.

Yumasai · 2025-12-13 21:45:19

6.5.4-3

Name            : kwin
Version         : 6.5.4-3
Description     : An easy to use, but flexible, Wayland compositor
Architecture    : x86_64
URL             : https://kde.org/plasma-desktop/
Licenses        : LGPL-2.0-or-later
Groups          : plasma
Provides        : None
Depends On      : aurorae  breeze  gcc-libs  glibc  plasma-activities  kauth  kcmutils  kcolorscheme  kconfig  kcoreaddons  kcrash  kdbusaddons  kdeclarative  kdecoration  kglobalaccel
                  kglobalacceld  kguiaddons  ki18n  kidletime  kirigami  kitemmodels  knewstuff  knighttime  knotifications  kpackage  kquickcharts  kscreenlocker  kservice  ksvg
                  kwayland  kwidgetsaddons  kwindowsystem  kxmlgui  lcms2  libcanberra  libdisplay-info  libdrm  libei  libepoxy  libinput  libpipewire  libqaccessibilityclient-qt6
                  libxcb  libxcvt  libxkbcommon  mesa  pipewire-session-manager  libplasma  qt6-5compat  qt6-base  qt6-declarative  qt6-sensors  qt6-svg  qt6-tools  systemd-libs
                  wayland  xcb-util-keysyms  xcb-util-wm
Optional Deps   : plasma-keyboard: virtual keyboard
Required By     : kdeplasma-addons  plasma-workspace
Optional For    : None
Conflicts With  : None
Replaces        : None
Installed Size  : 29.74 MiB
Packager        : Antonio Rojas <arojas@archlinux.org>
Build Date      : Fri 12 Dec 2025 04:25:28 PM CET
Install Date    : Fri 12 Dec 2025 04:58:42 PM CET
Install Reason  : Explicitly installed
Install Script  : Yes
Validated By    : Signature

seth · 2025-12-13 21:47:25

That "unfortunately" has the supposed freeze fix, you've had such freeze after the update (and restarting kwin) yesterday?

Yumasai · 2025-12-13 21:49:16

Yeah because as you can see here, I was still using the 6.5.4-3 wayland compositor "with the fix" there on wayland, when it occured. (a system boot [kernel loading] can be seen at the initial top line, so yes, rebooted too)

Last edited by Yumasai (2025-12-13 21:50:21)

Arch Linux

#1 2025-12-13 13:21:46