KDE Plasma freezes everything, then restarts after a few minutes (OOM)

Big Scorpio · 2025-02-01 01:44:23

Hello Arch Friends,

I am encountering this problem since I've switched from X11 to Wayland. There is an issue with my system, that it randomly freezes up everything, videos, games, sound, mouse movements is hindered and it's as if the machine's FPS has decreased to 1-3 seconds per frame (PER FRAME!). Then Plasmashell closes and then reopens and my system resumes operations.

I captured a system journalctl output here -> http://0x0.st/88Dt.txt

What seems suspicious to me are these:

// this one looks to me as the related side effect of the system freezing up
Feb 01 02:29:34 Nexus kwin_wayland[947]: kwin_libinput: Libinput: event3  - USB OPTICAL MOUSE : client bug: event processing lagging behind by 44ms, your system is too slow

// OOM indication
Feb 01 02:30:06 Nexus kernel: lowmem_reserve[]: 0 0 0 0 0
Feb 01 02:30:06 Nexus kernel: Node 0 DMA: 4*4kB (M) 4*8kB (M) 3*16kB (M) 1*32kB (M) 0*64kB 0*128kB 0*256kB 0*512kB 0*1024kB 0*2048kB 0*4096kB = 128kB
Feb 01 02:30:06 Nexus kernel: Node 0 DMA32: 886*4kB (UME) 239*8kB (UME) 148*16kB (UME) 85*32kB (UME) 31*64kB (UME) 30*128kB (UM) 3*256kB (U) 2*512kB (M) 0*1024kB 1*2048kB (M) 0*4096kB = 20208kB
Feb 01 02:30:06 Nexus kernel: Node 0 Normal: 1242*4kB (UME) 757*8kB (UME) 468*16kB (UME) 232*32kB (UME) 115*64kB (UME) 48*128kB (UM) 18*256kB (UM) 13*512kB (UM) 0*1024kB 0*2048kB 0*4096kB = 50704kB
Feb 01 02:30:06 Nexus kernel: Node 0 hugepages_total=0 hugepages_free=0 hugepages_surp=0 hugepages_size=1048576kB
Feb 01 02:30:06 Nexus kernel: Node 0 hugepages_total=0 hugepages_free=0 hugepages_surp=0 hugepages_size=2048kB
Feb 01 02:30:06 Nexus kernel: 18068 total pagecache pages
Feb 01 02:30:06 Nexus kernel: 5011 pages in swap cache
Feb 01 02:30:06 Nexus kernel: Free swap  = 144kB
Feb 01 02:30:06 Nexus kernel: Total swap = 12391928kB
Feb 01 02:30:06 Nexus kernel: 2066492 pages RAM
Feb 01 02:30:06 Nexus kernel: 0 pages HighMem/MovableOnly
Feb 01 02:30:06 Nexus kernel: 64674 pages reserved
Feb 01 02:30:06 Nexus kernel: 0 pages cma reserved
Feb 01 02:30:06 Nexus kernel: 0 pages hwpoisoned

// the thing showing, that it stopped plasmashell
Out of memory: Killed process 1122 (plasmashell) total-vm:8597212kB, anon-rss:844348kB, file-rss:20936kB, shmem-rss:0kB, UID:1000 pgtables:11384kB oom_score_adj:200

Is Wayland really that RAM hungry or what's going on? What can I do? I currently have 8 GB in RAM and I remember setting swap size to 8 GB too, yet I see the swap is bigger than what I've allocated when i installed arch linux.

12391928 kB ÷ 1024 bytes = 12100.5 MB ÷ 1024 = 11.81 GB

Last edited by Big Scorpio (2025-02-01 01:45:32)

ewaller · 2025-02-01 16:02:46

Wayland does not use all that much RAM. With no display manger, running Hyprland, a kitty terminal running bash this system uses 2.4 GiB.

KDE is a different animal. So are video and games. It sounds like your system is thrashing. Reboot, and before you start opening stuff, run the command

free -h

and post the results here. It will tell us what is your baseline and what swap is available.

Last edited by ewaller (2025-02-01 16:03:35)

Big Scorpio · 2025-02-01 18:03:07

               total        used        free      shared  buff/cache   available
Mem:           7.6Gi       3.0Gi       2.9Gi       390Mi       2.4Gi       4.6Gi
Swap:           11Gi          0B        11Gi

what's next?

ewaller · 2025-02-01 19:28:41

So, your baseline RAM is about 3 GiB, and you do have 11 GiB of swap space. I assume the system is running reasonably well at this point.
Try something and watch the output of that command. See what starts to use memory. Especially see what causes you to start using swap. You might use a tool such as htop to track memory use on a program by program basis. My bet is that something is leaking RAM. Allocating, using it, but forgetting to release it and then allocating more.

Edit: But we do now now your system uses more about a third of your memory right out of the gate.

Last edited by ewaller (2025-02-01 19:29:51)

Big Scorpio · 2025-02-01 22:45:34

Got an intermediate report so far. I actually went and upgraded my RAM to 32 GB. It was due already.
Anyway, the stall happened again and htop reports, that plasmashell occupies the most RAM of all. It may be limited information so far as to why plasmashell eats up RAM and becomes unresponsive. I wasn't able to take screenshots (so i took a photo), launch VLC or open the plasmashell menu.

Here are the images with htop showing RAM usage. Before terminating plasmashell https://i.imgur.com/B2Dtlvq.jpeg and after terminating plasmashell https://i.imgur.com/7SygyoD.png

According to the first image, the second most program using RAM is AI2U With you til the end, a game on steam. But this hiccup happens even outside of playing.

UPDATE:
I captured sysjournalctl too. The logs look concerning. See http://0x0.st/8Ksh.txt

What do i do next?

Last edited by Big Scorpio (2025-02-01 23:22:30)

seth · 2025-02-02 21:31:51

https://bbs.archlinux.org/viewtopic.php?id=303076 ?
(It's nvidia)

Big Scorpio · 2025-02-03 15:00:59

seth wrote:

https://bbs.archlinux.org/viewtopic.php?id=303076 ?
(It's nvidia)

That's cool and informative seth pal, but what action shall i attempt to undertake? Downgrade Nvidia? disable this xfwm thingy? Wait until the next pacman -Syu becomes available (oh hell no)? I can confirm I too have version 565.77 just like in the bug report found in the thread.

[motherfucker_big_scorpio@Nexus ~]$ ps aux | grep xfwm
motherf+    3500  0.0  0.0   7380  2116 pts/1    S+   12:24   0:00 grep --color=auto xfwm
[motherfucker_big_scorpio@Nexus ~]$ ps aux | grep xfw
motherf+    3505  0.0  0.0   7380  2196 pts/1    S+   12:24   0:00 grep --color=auto xfw
[motherfucker_big_scorpio@Nexus ~]$ systemctl | grep xfw
[motherfucker_big_scorpio@Nexus ~]$ systemctl | grep nvidia
[motherfucker_big_scorpio@Nexus ~]$ systemctl | grep nvi
[motherfucker_big_scorpio@Nexus ~]$ systemctl | grep sddm
  sddm.service                                                                                                                       loaded active running   Simple Desktop Display Manager
[motherfucker_big_scorpio@Nexus ~]$ systemctl | grep xf
[motherfucker_big_scorpio@Nexus ~]$ systemctl | grep ram
  sys-devices-virtual-block-zram0.device                                                                                             loaded active plugged   /sys/devices/virtual/block/zram0
  systemd-zram-setup@zram0.service                                                                                                   loaded active exited    Create swap on /dev/zram0
  system-systemd\x2dzram\x2dsetup.slice                                                                                              loaded active active    Slice /system/systemd-zram-setup
  dev-zram0.swap                                                                                                                     loaded active active    Compressed Swap on /dev/zram0
[motherfucker_big_scorpio@Nexus ~]$

Today it happened again. It shamelessly ate up all my 32 gigabyte RAM and 11 GB swap.

Edit:
I'll try downgrading nvidia. it seems this is the obvious way to avoid this problem.

Last edited by Big Scorpio (2025-02-03 15:06:37)

seth · 2025-02-03 15:10:13

downgrading nvidia […] seems […] the obvious way to avoid this problem.

xfce isn't relevant here, I just referenced the thread as "this is a known issue and likely not plasma but nvidia related"

If the older driver doesn't fix it (nb you'll have to use the dkms package! don#t just downgrade the driver in isolation) we'd have to look for other causes.

Big Scorpio · 2025-02-04 13:22:44

I require further assistance please. I intend to downgrade nvidia.

I did perform pacman -Syu for February this year to see if the problem was patched. The RAM clogging part has reoccurred. Here's the nvidia version (not the -dkms version)

[motherfucker_big_scorpio@Nexus ~]$ pacman -Q nvidia
nvidia 570.86.16-2

Downgrading nvidia was not successful for me. Something was always getting in the way as i followed down the rabbit hole.

[motherfucker_big_scorpio@Nexus ~]$ sudo downgrade nvidia
[sudo] password for motherfucker_big_scorpio: 
:: Retrieving packages...
 nvidia-560.35.03-1-x86_64                                         51.4 MiB  7.49 MiB/s 00:07 [#######################################################] 100%
loading packages...
warning: downgrading package nvidia (570.86.16-2 => 560.35.03-1)
resolving dependencies...
warning: cannot resolve "nvidia-utils=560.35.03", a dependency of "nvidia"
:: The following package cannot be upgraded due to unresolvable dependencies:
      nvidia

:: Do you want to skip the above package for this upgrade? [y/N] 
error: failed to prepare transaction (could not satisfy dependencies)
:: unable to satisfy dependency 'nvidia-utils=560.35.03' required by nvidia
[motherfucker_big_scorpio@Nexus ~]$ 
[motherfucker_big_scorpio@Nexus ~]$ sudo downgrade nvidia
loading packages...
warning: downgrading package nvidia (570.86.16-2 => 560.35.03-1)
resolving dependencies...
warning: cannot resolve "nvidia-utils=560.35.03", a dependency of "nvidia"
:: The following package cannot be upgraded due to unresolvable dependencies:
      nvidia

:: Do you want to skip the above package for this upgrade? [y/N] y
looking for conflicting packages...
 there is nothing to do
add nvidia to IgnorePkg? [y/N] y
[motherfucker_big_scorpio@Nexus ~]$ 
[motherfucker_big_scorpio@Nexus ~]$ sudo downgrade nvidia nvidia-utils
:: Retrieving packages...
 nvidia-utils-560.35.03-1-x86_64                                  229.4 MiB  8.04 MiB/s 00:29 [#######################################################] 100%
loading packages...
warning: downgrading package nvidia (570.86.16-2 => 560.35.03-1)
warning: downgrading package nvidia-utils (570.86.16-2 => 560.35.03-1)
resolving dependencies...
looking for conflicting packages...
error: failed to prepare transaction (could not satisfy dependencies)
:: installing nvidia-utils (560.35.03-1) breaks dependency 'nvidia-utils=570.86.16' required by lib32-nvidia-utils
[motherfucker_big_scorpio@Nexus ~]$ 
[motherfucker_big_scorpio@Nexus ~]$ sudo downgrade nvidia nvidia-utils lib32-nvidia-utils
[sudo] password for motherfucker_big_scorpio: 
:: Retrieving packages...
 lib32-nvidia-utils-560.35.03-1-x86_64                             39.1 MiB  6.78 MiB/s 00:06 [#######################################################] 100%
loading packages...
warning: downgrading package nvidia (570.86.16-2 => 560.35.03-1)
warning: downgrading package nvidia-utils (570.86.16-2 => 560.35.03-1)
warning: downgrading package lib32-nvidia-utils (570.86.16-1 => 560.35.03-1)
resolving dependencies...
looking for conflicting packages...

Packages (3) lib32-nvidia-utils-560.35.03-1  nvidia-560.35.03-1  nvidia-utils-560.35.03-1

Total Installed Size:   889.86 MiB
Net Upgrade Size:      -216.59 MiB

:: Proceed with installation? [Y/n] 
(3/3) checking keys in keyring                                                                [#######################################################] 100%
(3/3) checking package integrity                                                              [#######################################################] 100%
(3/3) loading package files                                                                   [#######################################################] 100%
(3/3) checking for file conflicts                                                             [#######################################################] 100%
error: failed to commit transaction (conflicting files)
nvidia-utils: /usr/lib/libnvidia-egl-gbm.so exists in filesystem (owned by egl-gbm)
nvidia-utils: /usr/lib/libnvidia-egl-gbm.so.1 exists in filesystem (owned by egl-gbm)
nvidia-utils: /usr/share/egl/egl_external_platform.d/15_nvidia_gbm.json exists in filesystem (owned by egl-gbm)
Errors occurred, no packages were upgraded.
[motherfucker_big_scorpio@Nexus ~]$ 
[motherfucker_big_scorpio@Nexus ~]$ pacman -Qi egl-gbm
Name            : egl-gbm
Version         : 1.1.2-1
Description     : The GBM EGL external platform library
Architecture    : x86_64
URL             : https://github.com/NVIDIA/egl-gbm
Licenses        : MIT
Groups          : None
Provides        : libnvidia-egl-gbm.so=1
Depends On      : eglexternalplatform  glibc  mesa  libdrm
Optional Deps   : None
Required By     : nvidia-utils
Optional For    : None
Conflicts With  : None
Replaces        : None
Installed Size  : 27.47 KiB
Packager        : Sven-Hendrik Haase <svenstaro@archlinux.org>
Build Date      : Thu Sep 19 09:32:24 2024
Install Date    : Mon Dec 30 18:53:27 2024
Install Reason  : Installed as a dependency for another package
Install Script  : No
Validated By    : Signature

[motherfucker_big_scorpio@Nexus ~]$ 
[motherfucker_big_scorpio@Nexus ~]$ sudo pacman -R egl-gbm
[sudo] password for motherfucker_big_scorpio: 
checking dependencies...
error: failed to prepare transaction (could not satisfy dependencies)
:: removing egl-gbm breaks dependency 'egl-gbm' required by nvidia-utils
[motherfucker_big_scorpio@Nexus ~]$ 
[motherfucker_big_scorpio@Nexus ~]$ sudo downgrade egl-gbm
Invalid choice
Unable to downgrade egl-gbm
[motherfucker_big_scorpio@Nexus ~]$

Noob question now, how do I really downgrade nvidia? Seth, were you saying I should start using nvidia-dkms and not nvidia? How do I do that? chroot into my arch linux, uninstall nvidia (with pacman) and then install nvidia-dkms ? A specific version of it? Is this OOM problem happening only on nvidia or nvidia-dkms too?

seth · 2025-02-04 13:43:25

I should start using nvidia-dkms and not nvidia?

https://wiki.archlinux.org/title/Dynami … le_Support
Install the nvidia-dkms package (you will have to downgrade nvidida-utils and lib32-nvidida-utils along the nvidia-package regardless!) and the linux-headers package (resp. the one relevant for your kernel(s))
You'll also have to remove egl-gbm which at some point was split off nvidia-utils and collides with it.

The OOM problem is, iff, bound to the drvier version - regardless whether you're using the precompiled package in the repos or the dkms variant.

Big Scorpio · 2025-02-04 14:43:36

seth wrote:

I should start using nvidia-dkms and not nvidia?
https://wiki.archlinux.org/title/Dynami … le_Support
Install the nvidia-dkms package (you will have to downgrade nvidida-utils and lib32-nvidida-utils along the nvidia-package regardless!) and the linux-headers package (resp. the one relevant for your kernel(s))
You'll also have to remove egl-gbm which at some point was split off nvidia-utils and collides with it.
The OOM problem is, iff, bound to the drvier version - regardless whether you're using the precompiled package in the repos or the dkms variant.

Complied. I installed "dkms" and "linux-headers" (I'm using standard linux kernel, not LTS or anything). Before that however I (had to) forcibly delete egl-gbm like so "sudo pacman -Rdd egl-gbm". Then i downgraded the 3 nvidia packages like so "sudo downgrade nvidia-utils lib32-nvidia-utils nvidia". Added all 3 to IgnorePkg.

Did i do it right? To me it seems right, now

[motherfucker_big_scorpio@Nexus ~]$ pacman -Q nvidia
nvidia 560.35.03-1

What will dkms and linux-headers do from here? Will they do something after having them only installed on my system or need I do anything else from here on?
Any major risks or bad practices, that I put Nvidia on IgnorePkg and use a fixed version before this memory hogging started to occur?

seth · 2025-02-04 14:59:07

Everytime you update the kernel, dkms will automatically build a matching nvidia driver of the installed (560.35) version.
W/ kernel API changes (mainly GPL restrictions) the (older) nvidia modules might at some point fail to build in which case you'll block nouveau (via nvidia-utils) but don't have an nvidia kernel module.
You'll at this point have to look for nvidia patches, stall the kernel as well or move to eg. the LTS kernel.

Edit: according to https://bbs.archlinux.org/viewtopic.php … 9#p2224549 you'll have to use the LTS kernel (and linux-lts-headers) anyway?
"dkms status" will show the driver only as added, but not installed? The build failed?

Last edited by seth (2025-02-04 15:00:57)

Big Scorpio · 2025-02-04 15:47:47

[motherfucker_big_scorpio@Nexus ~]$ dkms status
[motherfucker_big_scorpio@Nexus ~]$

Maybe. This is new territory for me.
Shall I switch to linux-lts i wouldn't know how to. Can I do that on my system logged in or should it be done via chroot-ing into it?

V1del · 2025-02-04 15:55:34

You can just install linux-lts{-headers} even alongside the normal linux package then you'll have to configure your bootloader to boot into that.

And for dkms status to work properly you need to install nvidia-dkms 560.35.03 after which every kernel update will rebuild a relevant nvidia module for all kernels that have the -headers installed

Big Scorpio · 2025-02-04 21:32:42

Man, i'm doing things the hard way the way it goes. I ended up with a broken sddm and spent an hour fixing it by installing the latest nvidia crap (and then replacing it with nvidia-dkms).

Anyway, I downgraded to nvidia-dkms 560.35.03 just now. Let's see how it goes. Here's the results so far:

(6/8) Install DKMS modules
==> dkms install --no-depmod nvidia/560.35.03 -k 6.13.1-arch1-1

Error! Bad return status for module build on kernel: 6.13.1-arch1-1 (x86_64)
Consult /var/lib/dkms/nvidia/560.35.03/build/make.log for more information.
==> WARNING: `dkms install --no-depmod nvidia/560.35.03 -k 6.13.1-arch1-1' exited 10

]
==> WARNING: No fsck helpers found. fsck will not be run on boot.
==> ERROR: module not found: 'nvidia'
==> Generating module dependencies
==> Creating zstd-compressed initcpio image: '/boot/initramfs-linux.img'
==> WARNING: errors were encountered during the build. The image may not be complete.
==> Building image from preset: /etc/mkinitcpio.d/linux.preset: 'fallback'
==> Using default configuration file: '/etc/mkinitcpio.conf'
  -> -k /boot/vmlinuz-linux -g /boot/initramfs-linux-fallback.img -S autodetect
==> Starting build: '6.13.1-arch1-1'
  -> Running build hook: [base]


==> WARNING: Possibly missing firmware for module: 'qat_420xx'
  -> Running build hook: [lvm2]
  -> Running build hook: [filesystems]
  -> Running build hook: [fsck]
==> ERROR: module not found: 'nvidia'
==> Generating module dependencies
==> Creating zstd-compressed initcpio image: '/boot/initramfs-linux-fallback.img'
==> WARNING: errors were encountered during the build. The image may not be complete.
error: command failed to execute correctly
(8/8) Reloading system bus configuration...
add nvidia-dkms to IgnorePkg? [y/N] y
[motherfucker_big_scorpio@Nexus ~]$ 
[motherfucker_big_scorpio@Nexus ~]$ 
[motherfucker_big_scorpio@Nexus ~]$ pacman -Q nvidia
nvidia-dkms 560.35.03-1
[motherfucker_big_scorpio@Nexus ~]$ pacman -Q nvidia-utils
nvidia-utils 560.35.03-1
[motherfucker_big_scorpio@Nexus ~]$ pacman -Q lib32-nvidia-utils
lib32-nvidia-utils 560.35.03-1
[motherfucker_big_scorpio@Nexus ~]$ pacman -Q linux
linux 6.13.1.arch1-1
[motherfucker_big_scorpio@Nexus ~]$ dkms status
nvidia/560.35.03: added
[motherfucker_big_scorpio@Nexus ~]$

Does this look good?

From the make logs of "cat /var/lib/dkms/nvidia/560.35.03/build/make.log"

  CC [M]  nvidia-modeset/nv-kthread-q.o
  CC [M]  nvidia-drm/nvidia-drm.o
  CC [M]  nvidia-drm/nvidia-drm-drv.o
nvidia-modeset/nvidia-modeset-linux.c:1087:24: warning: no previous prototype for ‘nvkms_open_common’ [-Wmissing-prototypes]
 1087 | struct nvkms_per_open *nvkms_open_common(enum NvKmsClientType type,
      |                        ^~~~~~~~~~~~~~~~~
nvidia-modeset/nvidia-modeset-linux.c:1139:6: warning: no previous prototype for ‘nvkms_close_pm_locked’ [-Wmissing-prototypes]
 1139 | void nvkms_close_pm_locked(struct nvkms_per_open *popen)
      |      ^~~~~~~~~~~~~~~~~~~~~
nvidia-modeset/nvidia-modeset-linux.c:1202:5: warning: no previous prototype for ‘nvkms_ioctl_common’ [-Wmissing-prototypes]
 1202 | int nvkms_ioctl_common
      |     ^~~~~~~~~~~~~~~~~~
  CC [M]  nvidia-drm/nvidia-drm-utils.o
  CC [M]  nvidia-drm/nvidia-drm-crtc.o
  CC [M]  nvidia-drm/nvidia-drm-encoder.o
  CC [M]  nvidia-drm/nvidia-drm-connector.o
  CC [M]  nvidia-drm/nvidia-drm-gem.o
  CC [M]  nvidia-drm/nvidia-drm-fb.o
nvidia-drm/nvidia-drm-drv.c:207:6: error: ‘const struct drm_mode_config_funcs’ has no member named ‘output_poll_changed’
  207 |     .output_poll_changed = nv_drm_output_poll_changed,
      |      ^~~~~~~~~~~~~~~~~~~
nvidia-drm/nvidia-drm-drv.c:207:28: error: initialization of ‘struct drm_atomic_state * (*)(struct drm_device *)’ from incompatible pointer type ‘void (*)(struct drm_device *)’ [-Wincompatible-pointer-types]
  207 |     .output_poll_changed = nv_drm_output_poll_changed,
      |                            ^~~~~~~~~~~~~~~~~~~~~~~~~~
nvidia-drm/nvidia-drm-drv.c:207:28: note: (near initialization for ‘nv_mode_config_funcs.atomic_state_alloc’)
make[4]: *** [/usr/lib/modules/6.13.1-arch1-1/build/scripts/Makefile.build:194: nvidia-drm/nvidia-drm-drv.o] Error 1
make[4]: *** Waiting for unfinished jobs....
  CC [M]  nvidia-drm/nvidia-drm-modeset.o
make[3]: *** [/usr/lib/modules/6.13.1-arch1-1/build/Makefile:1989: .] Error 2
make[2]: *** [/usr/lib/modules/6.13.1-arch1-1/build/Makefile:251: __sub-make] Error 2
make[2]: Leaving directory '/var/lib/dkms/nvidia/560.35.03/build'
make[1]: *** [Makefile:251: __sub-make] Error 2
make[1]: Leaving directory '/usr/lib/modules/6.13.1-arch1-1/build'
make: *** [Makefile:89: modules] Error 2

Last edited by Big Scorpio (2025-02-04 21:34:45)

seth · 2025-02-04 21:34:30

No, as mentioned that version does not build on 6.13 - you want to try/use the LTS kernel.

Big Scorpio · 2025-02-04 21:58:19

Hm, OK. I understand.
If I understood correctly, using linux 6.11.9 ought to work, please confirm for my (and others) information. But i complied with your suggestion because i want to get this over with.

[motherfucker_big_scorpio@Nexus ~]$ sudo pacman -Q linux-lts
linux-lts 6.12.12-1
[motherfucker_big_scorpio@Nexus ~]$ sudo pacman -Q linux-lts-headers
linux-lts-headers 6.12.12-1

I used grub-customizer to select the linux-lts kernel to boot with (Arch Linux, with Linux linux-lts) (Advanced options for Arch Linux>Arch Linux, with Linux linux-lts - being my default predefined entry). I didn't fully grasp the instructions to do that manually.
Am I good now?

Edit:

But since kernel 6.12 can't be built against nvidia 560.35.03-19 I had to freeze kernel updates too.

Grrr.... i think i know what to do next. Downgrade linux-lts to 6.6.72 or use linux 6.11.9 (?)

Edit 2:

Packages (2) linux-6.11.9.arch1-1  linux-headers-6.11.9.arch1-1

(3/4) Install DKMS modules
==> dkms install --no-depmod nvidia/560.35.03 -k 6.11.9-arch1-1
==> depmod 6.11.9-arch1-1
(4/4) Updating linux initcpios...
==> Building image from preset: /etc/mkinitcpio.d/linux.preset: 'default'
==> Using default configuration file: '/etc/mkinitcpio.conf'
  -> -k /boot/vmlinuz-linux -g /boot/initramfs-linux.img
==> Starting build: '6.11.9-arch1-1'
  -> Running build hook: [base]

(then comes the mkinitcpio.conf part. it ends with)

==> Generating module dependencies
==> Creating zstd-compressed initcpio image: '/boot/initramfs-linux-fallback.img'
  -> Early uncompressed CPIO image generation successful
==> Initcpio image generation successful
error: command failed to execute correctly
add linux to IgnorePkg? [y/N] y
add linux-headers to IgnorePkg? [y/N] y
[motherfucker_big_scorpio@Nexus ~]$ 
[motherfucker_big_scorpio@Nexus ~]$ pacman -Q linux
linux 6.11.9.arch1-1
[motherfucker_big_scorpio@Nexus ~]$ pacman -Q linux-headers
linux-headers 6.11.9.arch1-1
[motherfucker_big_scorpio@Nexus ~]$

Man, i hope i'm in the clear now.

Last edited by Big Scorpio (2025-02-04 22:07:18)

seth · 2025-02-04 22:27:14

"dkms status" needs to say "installed" for the relevant kernel.
Sorry, I didn't relalize that LTS had recently moved to 6.12…

Big Scorpio · 2025-02-04 23:02:51

Right you are, Seth.

[motherfucker_big_scorpio@Nexus ~]$ dkms status
nvidia/560.35.03, 6.11.9-arch1-1, x86_64: installed
[motherfucker_big_scorpio@Nexus ~]$

Seems successful. But before we mark this as SOLVED, i'd like to go a few days like so to confirm the fix. I'll do it myself (if i won't die before that).

Big Scorpio · 2025-02-05 18:21:36

Not solved unfortunately. It happened again. See here http://0x0.st/8KCR.txt
Ignore the firefox logs. I attempted to enable vaapi (Intel 6700K, GTX 1080 Ti) through about:config and found it did nothing, so i disabled it afterwards.

I am thinking of downgrading plasma itself somehow.
Any ideas what to do next?

Big Scorpio · 2025-02-06 09:49:04

Downgrading plasma-desktop to 6.2.4-1 (the one I had since the beginning and where this problem was not prevalent) did not work either. plasmashell still eats up all RAM and stalls the system.
Do I really have to nuke the whole damn system?

seth · 2025-02-06 14:58:37

This will hardly achieve anything if the problem is with some desktop gadget and the config in your $HOME

1st off all, does plasmashell on X11 behave the same?
2nd: do you use any kind of (3rd party) plasmoids?
3. Do other (GL heavy) processes leak RAM?

Big Scorpio · 2025-02-06 19:07:24

1. I don't know. I don't want to use X11 due to it's unnerving sluggishness here (and i don't know what to try next in that context) -> https://bbs.archlinux.org/viewtopic.php?id=302113&p=2
2. No, i don't think I did anything to use a 3rd party plasmoid.
3. How can I find out? GL as in OpenGL? Me novice with that sector.

See here for screenshots of my next interesting update on this issue -> https://imgur.com/a/v01wNnj

I noticed how plasmashell consumes as much as 50% of my CPU too (wtf for?).
I did a "sterilization" test with KDE plasma being the focus by quitting everything except OBS and htop and then resetting the plasma config under $HOME/.config - there was a file called "plasma-org.kde.plasma.desktop-appletsrc" that i renamed to something else and performed a kill -9 PID of plasmashell. I found out about it by browsing online to see who else has the same problem as I have.

That's where the second screenshot comes in. less than 2 GB of RAM being occupied. Something with plasma desktop config was causing this RAM occupation invasion. The freezing up and restarting of plasmashell did not occur after.

Switched back to my original "plasma-org.kde.plasma.desktop-appletsrc" and the RAM occupation and freeze up happened again. Interestingly this happened while i let the desktop wallpaper being visible and not covered with any program windows. My desktop wallpaper was a GIF. I changed the wallpaper to something, that isn't a GIF and since then plasmashell was behaving OK, no freezing up or RAM occupation.

It really looks to me like KDE Plasma Desktop has a problem with Gif wallpapers eating up my CPU and overclogging my RAM. Because when I switched up the wallpaper, the system kept running just fine and dandy.

Really now? It appears to be a KDE problem above all. I might file a bug on their premises.
For now I'll keep this topic open until i can confidently confirm it was indeed the Gif alone doing this bad behavior.

Edit:

Aha, i see someone reported this exact problem, that i had with the gif wallpaper memory leak in KDE forum and apparently filed a bug on KDE with no reference to the filed bug: https://discuss.kde.org/t/major-memory-leak/20394

Last edited by Big Scorpio (2025-02-07 01:28:43)

irarch · 2025-02-07 07:37:00

Rather than downgrading have you tried updating the driver?

The current version of the nvidia package is 570.86.16-2

seth · 2025-02-07 09:00:31

I changed the wallpaper to something, that isn't a GIF and since then plasmashell was behaving OK, no freezing up or RAM occupation.

bug on KDE with no reference to the filed bug: https://discuss.kde.org/t/major-memory-leak/20394

Please always remember to mark resolved threads by editing your initial posts subject - so others will know that there's no task left, but maybe a solution to find.
Thanks.
Also adjust the subject to hightlight the cause.

Arch Linux

#1 2025-02-01 01:44:23

KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#2 2025-02-01 16:02:46

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#3 2025-02-01 18:03:07

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#4 2025-02-01 19:28:41

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#5 2025-02-01 22:45:34

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#6 2025-02-02 21:31:51

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#7 2025-02-03 15:00:59

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#8 2025-02-03 15:10:13

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#9 2025-02-04 13:22:44

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#10 2025-02-04 13:43:25

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#11 2025-02-04 14:43:36

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#12 2025-02-04 14:59:07

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#13 2025-02-04 15:47:47

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#14 2025-02-04 15:55:34

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#15 2025-02-04 21:32:42

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#16 2025-02-04 21:34:30

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#17 2025-02-04 21:58:19

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#18 2025-02-04 22:27:14

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#19 2025-02-04 23:02:51

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#20 2025-02-05 18:21:36

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#21 2025-02-06 09:49:04

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#22 2025-02-06 14:58:37

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#23 2025-02-06 19:07:24

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#24 2025-02-07 07:37:00

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

#25 2025-02-07 09:00:31

Re: KDE Plasma freezes everything, then restarts after a few minutes (OOM)

Board footer