[SOLVED] No video signal after system update & log in - Plasma Wayland

Big Scorpio · 2026-03-23 21:11:16

Readers note: In my case my video signal loss happened exclusively on KDE Plasma Wayland as my Desktop Environment, thus this is a problem, that is a mixture of System Admin fault (environment variables) and KDE Plasma on Wayland (not X11!).
TL-DR solution: in my "/etc/environment" file the environment variable KWIN_DRM_DEVICES was set to

KWIN_DRM_DEVICES="/dev/dri/card1:/dev/dri/card0"

Which is an unreliable way to do DRM device specification since the "/dev/dri/card*" items are not statically persistent and may change. This caused my video signal loss. Instead it had to be set to a persistent PCI HW path - for me it was "/dev/dri/by-path/pci-0000\:01\:00.0-card:/dev/dri/by-path/pci-0000\:00\:02.0-card". Yours may (and will) differ. See further down below how to (probably) fix yours like I did.

---

Hello Arch Friends,

So the system update has broken my Linux usability almost completely. Right now I'm sitting on Win 10, that I set up for dual booting for cases just like this.

Last I recall, I was using Linux with Kernel 6.12 LTS just fine. The regular rolling release kernel 6.18 or something (that was before the 6.18 LTS kernel was released) was exhibiting this exact problem too. Now after the system update I can use neither kernel, be it 6.18 LTS or the currently newer regular rolling release kernel at 6.19 as the video signal drops entirely right after logging in with my username on SDDM. I can't even activate other TTY's by pressing CTRL+ALT+F[1-4] in this state. Numlock and CAPS LOCK do function normally as they reflect their states on my keyboard, so it's not like my system became completely unresponsive. The only way out of there is by pressing on the power reset button on my chassis.

I'm using GRUB and KDE Plasma Wayland. Haven't tested it on other Desktop Environments yet.

I have a GTX 1080 Ti and have installed the nvidia-580xx versions of drivers as indicated by my earlier post here https://bbs.archlinux.org/viewtopic.php?id=312770 - which caused this problem as part of a side effect of solving my initial pacman/AUR issue. I'd have to think of a way to get the nvidia installations data and post it here someway.

How can we find out the problem for the video signal being killed?

Last edited by Big Scorpio (2026-03-25 21:08:32)

seth · 2026-03-23 22:00:16

Logging into KDE?
https://wiki.archlinux.org/title/KDE#Un … lution_set

However and wrt your other thread please ctrl+alt+f3, login there and post your complete system journal for the boot:

sudo journalctl -b | curl -F 'file=@-' https://0x0.st

Big Scorpio · 2026-03-24 19:05:52

seth wrote:

Logging into KDE?

Indeed, Wayland that is.
Interestingly I found, that this video signal loss occurs on KDE Plasma Wayland as Desktop Environment. I just logged in through Cinnamon (X11) like normally and it's working just fine.

I found, that my ".xinitrc" file in my HOME dir was empty and according to the wiki's recommendation ( https://wiki.archlinux.org/title/Xorg#Troubleshooting ), I have deleted it.
Furthermore I have inspected "~/.local/share/kscreen/" but I don't have a "kscreen" in that path.
Also I was not able to locate "xorg.conf" under
/etc
/etc/X11
~/.local/share/xorg
/usr/share/X11
/usr/share -- no "xorg" dir

By the looks of it, it seems, that KDE Plasma broke itself in the update.

[motherfucker_big_scorpio@Nexus ~]$ sudo cat /sys/module/nvidia_drm/parameters/modeset
Y

Here is my journalctl just as instructed -> https://0x0.st/PfPy.txt

What do I do next?

Note: I found a few other users having the exact same problem as I do here ( https://forum.garudalinux.org/t/no-vide … sion/35941 ) and here ( https://bbs.archlinux.org/viewtopic.php?id=293571 - solution unclear - suggested, that setting the monitor refresh rates to 60hz and then unplugging and replugging the HDMI/DP cables worked for them (?))

Last edited by Big Scorpio (2026-03-24 19:24:35)

seth · 2026-03-24 20:35:31

The journal doesn't cover a KDE login, there's a shady USB device (seems some external hub? dock?)

Mar 24 19:13:27 Nexus kernel: usb usb2-port1: Cannot enable. Maybe the USB cable is bad?
Mar 24 19:13:31 Nexus kernel: usb usb2-port1: Cannot enable. Maybe the USB cable is bad?
Mar 24 19:13:35 Nexus kernel: usb usb2-port1: Cannot enable. Maybe the USB cable is bad?

Try to log into a KDE/X11 session and post a journal that covers a video loss (you can most likely still switch the VT to obtain it) and you probably also want to test whether the problem also affects a fresh user account.

Big Scorpio · 2026-03-24 21:19:51

seth wrote:

there's a shady USB device (seems some external hub? dock?)

Yeah, I got 2 external USB hubs connected. It seems they "inherited" the same USB problem as the USB port I udev blacklisted earlier. (I really need to throw away - recycle my Motherboard and get a new one /!\)

Historically however the bad USB messages have never been the cause of video signal loss - for me at least.

Here is the journalctl with Plasma X11 post login (it works) -> https://0x0.st/PfNp.txt
Had to install the X11 package for plasma separately as KDE intends to deprecate and ditch the X11 variant thereof.
So we have isolated the problem exclusively to KDE Plasma Wayland now.
I also confirmed, that this problem happens on a brand new user too. Could it be, that a certain dependency of Plasma Wayland is causing this problem? I recall, that upon performing "pacman -Syu" it asked me to replace a Qt package, that i don't remember the name of anymore.

[motherfucker_big_scorpio@Nexus home]$  pacman -Q | grep qt
appstream-qt 1.1.2-1
bluez-qt 6.24.0-2
layer-shell-qt 6.6.3-1
libaccounts-qt 1.17-2
libdbusmenu-lxqt 0.3.0-1
libfm-qt 2.3.1-2
liblxqt 2.3.0-1
libqaccessibilityclient-qt6 0.6.0-1
libqtxdg 4.3.0-4
lximage-qt 2.3.0-1
lxqt-about 2.3.0-1
lxqt-admin 2.3.0-1
lxqt-archiver 1.3.0-2
lxqt-config 2.3.1-2
lxqt-globalkeys 2.3.0-1
lxqt-menu-data 2.3.0-1
lxqt-notificationd 2.3.1-1
lxqt-openssh-askpass 2.3.0-1
lxqt-policykit 2.3.0-1
lxqt-qtplugin 2.3.0-3
lxqt-runner 2.3.0-1
lxqt-session 2.3.0-2
lxqt-sudo 2.3.0-1
lxqt-themes 2.3.0-1
lxqt-wayland-session 0.3.2-1
modemmanager-qt 6.24.0-2
networkmanager-qt 6.24.0-2
obconf-qt 0.16.6-1
pavucontrol-qt 2.3.0-1
pcmanfm-qt 2.3.0-2
phonon-qt6 4.12.0-6
phonon-qt6-vlc 0.12.0-6
polkit-qt6 0.200.0-2
poppler-qt6 26.03.0-1
pulseaudio-qt 1.8.1-1
qca-qt6 2.3.10-5
qt5-base 5.15.18+kde+r109-2
qt5-declarative 5.15.18+kde+r23-1
qt5-svg 5.15.18+kde+r5-1
qt5-translations 5.15.18-1
qt5-wayland 5.15.18+kde+r55-1
qt5-x11extras 5.15.18-1
qt6-5compat 6.10.2-1
qt6-base 6.10.2-1
qt6-declarative 6.10.2-1
qt6-imageformats 6.10.2-1
qt6-location 6.10.2-1
qt6-multimedia 6.10.2-1
qt6-multimedia-ffmpeg 6.10.2-1
qt6-positioning 6.10.2-1
qt6-quick3d 6.10.2-1
qt6-quicktimeline 6.10.2-1
qt6-sensors 6.10.2-1
qt6-shadertools 6.10.2-1
qt6-speech 6.10.2-1
qt6-svg 6.10.2-1
qt6-tools 6.10.2-3
qt6-translations 6.10.2-1
qt6-virtualkeyboard 6.10.2-1
qt6-wayland 6.10.2-1
qt6-webchannel 6.10.2-1
qt6-webengine 6.10.2-2
qt6-websockets 6.10.2-1
qt6-webview 6.10.2-1
qt6pas 6.2.10-3
qterminal 2.3.0-1
qtermwidget 2.3.0-2
qtkeychain-qt6 0.15.0-4
qtxdg-tools 4.3.0-1
vlc-gui-qt 3.0.22-1
xdg-desktop-portal-lxqt 1.3.0-2

Here are some more things I have attempted without success so far:

https://bbs.archlinux.org/viewtopic.php … 1#p2155191 - Deleting the entire [Kscreen] section from "~/.config/kdeglobals file"
https://discuss.kde.org/t/reset-all-pla … -line/9503 - Deleting plasma config files
https://www.reddit.com/r/kde/comments/v … et_plasma/ - Deleting plasma config files

Additionally, just in case it has any relevance, here is my /etc/environment content if it has anything to do with it. I haven't changed it prior to this problem.

[motherfucker_big_scorpio@Nexus home]$ cat /etc/environment 
#
# This file is parsed by pam_env module
#
# Syntax: simple "KEY=VAL" pairs on separate lines
#
KWIN_DRM_DEVICES="/dev/dri/card1:/dev/dri/card0"

# new entries
QT_QPA_PLATFORMTHEME="wayland;xcb"

GBM_BACKEND=nvidia-drm
__GL_VRR_ALLOWED=1
__GL_NGX_ALLOWED=1
__GLX_VENDOR_LIBRARY_NAME=nvidia
LIBVA_DRIVER_NAME=nvidia
#__NV_PRIME_RENDER_OFFLOAD=0
#WLR_NO_HARDWARE_CURSORS=1
QSG_RENDERER_LOOP=threaded
VDPAU_DRIVER=nvidia
#NVD_BACKEND=direct

# Proton settings
PROTON_ENABLE_WAYLAND=1
PROTON_ENABLE_HDR=1

Last edited by Big Scorpio (2026-03-24 21:23:48)

seth · 2026-03-24 21:34:45

So we have isolated the problem exclusively to KDE Plasma Wayland now … this problem happens on a brand new user too

So we shall see a journal covering the wayland failure.

Couple of things
1. fix your locale

Mar 24 22:03:46 Nexus steam[1346]: [error] cannot open locale definition file `en_DE': No such file or directory
Mar 24 22:03:46 Nexus steam[1346]: pv-locale-gen: Unable to generate locale en_DE.UTF-8: 4
Mar 24 22:03:46 Nexus steam[1346]: setlocale "en_DE.UTF-8": No such file or directory

"en_DE.UTF-8" isn't a thing, as much as german marketing experts would like it to be.

Then there's a metric shit-ton of coredumps in /var/lib/systemd/coredump/ (mostly baloo, it seems), you might want to clean that up because drkonqi is processing all of them - resulting in spam littering 95% of the journal

seth · 2026-03-24 21:36:06

KWIN_DRM_DEVICES="/dev/dri/card1:/dev/dri/card0"

Remove that, there's only one GPU?
Scratch that, there're two drm devices but only one VGA one (nvidia)

Mar 24 22:02:36 Nexus kernel: i915 0000:00:02.0: [drm] [ENCODER:105:DDI A/PHY A] failed to retrieve link info, disabling eDP
Mar 24 22:02:36 Nexus kernel: i915 0000:00:02.0: [drm] *ERROR* Failed to probe lspcon

fails - still maybe look into that value and esp. what card0 and card1 actually are (the order isn't really deterministic)

Last edited by seth (2026-03-24 21:38:03)

Big Scorpio · 2026-03-25 00:36:12

*groan* i can tell this is gonna be lots of "fun".

First about the DRM devices

[motherfucker_big_scorpio@Nexus ~]$ lspci -nn | grep -E 'VGA|3D'
01:00.0 VGA compatible controller [0300]: NVIDIA Corporation GP102 [GeForce GTX 1080 Ti] [10de:1b06] (rev a1)
[motherfucker_big_scorpio@Nexus ~]$ lspci -nn | grep 'HD Graphics'
00:02.0 Display controller [0380]: Intel Corporation Skylake-S GT2 [HD Graphics 530] [8086:1912] (rev 06)
[motherfucker_big_scorpio@Nexus ~]$ ls -l /sys/class/drm/card*/device
lrwxrwxrwx 1 root root 0 Mar 25 01:00 /sys/class/drm/card1-DP-1/device -> ../../card1
lrwxrwxrwx 1 root root 0 Mar 25 01:00 /sys/class/drm/card1-DP-2/device -> ../../card1
lrwxrwxrwx 1 root root 0 Mar 25 01:00 /sys/class/drm/card1-DP-3/device -> ../../card1
lrwxrwxrwx 1 root root 0 Mar 25 01:00 /sys/class/drm/card1-HDMI-A-1/device -> ../../card1
lrwxrwxrwx 1 root root 0 Mar 25 01:00 /sys/class/drm/card1-HDMI-A-2/device -> ../../card1
lrwxrwxrwx 1 root root 0 Mar 25 01:00 /sys/class/drm/card1/device -> ../../../0000:00:02.0
lrwxrwxrwx 1 root root 0 Mar 25 01:00 /sys/class/drm/card2-DP-4/device -> ../../card2
lrwxrwxrwx 1 root root 0 Mar 25 01:00 /sys/class/drm/card2-DP-5/device -> ../../card2
lrwxrwxrwx 1 root root 0 Mar 25 01:00 /sys/class/drm/card2-DP-6/device -> ../../card2
lrwxrwxrwx 1 root root 0 Mar 25 01:00 /sys/class/drm/card2-DVI-D-1/device -> ../../card2
lrwxrwxrwx 1 root root 0 Mar 25 01:00 /sys/class/drm/card2-HDMI-A-3/device -> ../../card2
lrwxrwxrwx 1 root root 0 Mar 25 01:00 /sys/class/drm/card2/device -> ../../../0000:01:00.0
[motherfucker_big_scorpio@Nexus ~]$ 
[motherfucker_big_scorpio@Nexus ~]$ ls /dev/dri/card*
/dev/dri/card1  /dev/dri/card2
[motherfucker_big_scorpio@Nexus ~]$

/sys/class/drm/card1/device -> ../../../0000:00:02.0 -- seems to correspond to my Intel CPU integrated GPU HD Graphics 530
/sys/class/drm/card2/device -> ../../../0000:01:00.0 -- seems to correspond to my discrete GPU NVIDIA GeForce GTX 1080 Ti

I don't know if my memory tricks me but i can see the KWIN_DRM_DEVICES environment value is wrong already (it shouldn't have been). I've set it to card2, then card1 now in "/etc/environment".

Next, about locale. I've heard it like 10 times now and wasn't able to determine where steam gets the en_DE from. Until I ran these commands

[motherfucker_big_scorpio@Nexus home]$ locale -a
C
C.utf8
POSIX
de_DE.utf8
en_AG
en_AG.utf8
en_US.utf8
pt_PT.utf8
[motherfucker_big_scorpio@Nexus home]$ 
[motherfucker_big_scorpio@Nexus home]$ localectl list-locales
C.UTF-8
de_DE.UTF-8
en_AG.UTF-8
en_US.UTF-8
pt_PT.UTF-8
[motherfucker_big_scorpio@Nexus home]$ 
[motherfucker_big_scorpio@Nexus home]$ cat /etc/locale.gen | grep -v "#"
de_DE.UTF-8 UTF-8  
en_AG UTF-8  
en_US.UTF-8 UTF-8  
pt_PT.UTF-8 UTF-8  
[motherfucker_big_scorpio@Nexus home]$ cat /etc/locale.conf | grep -v "#"
LANG=en_US.UTF-8
LC_CTYPE=en_AG.UTF-8
LC_NUMERIC=en_AG.UTF-8
LC_TIME=en_AG.UTF-8
LC_COLLATE=en_AG.UTF-8
LC_MONETARY=pt_PT.UTF-8
LC_MESSAGES=en_AG.UTF-8
LC_PAPER=en_AG.UTF-8
LC_NAME=en_AG.UTF-8
LC_ADDRESS=en_AG.UTF-8
LC_TELEPHONE=en_AG.UTF-8
LC_MEASUREMENT=en_AG.UTF-8
LC_IDENTIFICATION=en_AG.UTF-8
[motherfucker_big_scorpio@Nexus home]$ 
[motherfucker_big_scorpio@Nexus home]$ env | grep -E "^LC_|^LANG" | sort
LANG=C.UTF-8
LANGUAGE=
LC_ADDRESS=en_AG.UTF-8
LC_ALL=C.UTF-8
LC_COLLATE=en_AG.UTF-8
LC_CTYPE=en_AG.UTF-8
LC_IDENTIFICATION=en_AG.UTF-8
LC_MEASUREMENT=en_AG.UTF-8
LC_MESSAGES=en_AG.UTF-8
LC_MONETARY=pt_PT.UTF-8
LC_NAME=en_AG.UTF-8
LC_NUMERIC=en_DE.UTF-8 # <-------- *** DING DING DING ***
LC_PAPER=en_AG.UTF-8
LC_TELEPHONE=en_AG.UTF-8
LC_TIME=en_CH.UTF-8
[motherfucker_big_scorpio@Nexus home]$

How? JUST HOW?

So I'm trying some things now.

in my "/etc/environment"

# locale
LC_NUMERIC=C.UTF-8

in my ~/.bashrc

# locale
unset LC_NUMERIC
export LC_NUMERIC=de_DE:C

Here specifically I am hoping to chain fallback locales.

Let's see which one works - if at all. Will report next day.

V1del · 2026-03-25 01:17:09

card1 and card2 are never going to be deterministic and will potentially fail from boot to boot

Assuming you don't actually require the integrated GPU, I suggest you get rid of all of these special environment variables to try and coerce things to nvidia and simply disable the integrated GPU in your UEFI (or if that's not possible, blacklist i915)

If you actually do need it, you can also list the stable PCI identifiers as the KWIN_DRM devices: https://invent.kde.org/plasma/kwin/-/wi … rm_devices

Last edited by V1del (2026-03-25 01:21:10)

seth · 2026-03-25 07:50:04

https://wiki.archlinux.org/title/KDE#Pl … e_settings ?

Big Scorpio · 2026-03-25 20:51:36

V1del wrote:

If you actually do need it, you can also list the stable PCI identifiers as the KWIN_DRM devices: https://invent.kde.org/plasma/kwin/-/wi … rm_devices

There is a reason i deliberately have a hybrid GPU setup. I use the Nvidia one for multimedia and gaming and the Intel one (or AMD in the future) for recording.

Anyway, this environment variable change fixed it for me. I am back in KDE Plasma Wayland. Yay!

KWIN_DRM_DEVICES="/dev/dri/by-path/pci-0000\:01\:00.0-card:/dev/dri/by-path/pci-0000\:00\:02.0-card"

Also i was digging whether i really need the following environment variable for my hybrid GPU system use case. This is the best information i could dig up so far but still I rather ask experienced humans than AI LLMs.

# Graphics Buffer Manager (GBM) - use Direct Rendering Manager (DRM) with this backend. Needed for modern display servers (e.g. Wayland)
# Consider removing. May interfere with Intel or AMD capturing // Note: OBS allows selecting the GPU for encoding for video capturing and streaming manually, so this setting may not be relevant when using OBS
GBM_BACKEND=nvidia-drm

# Force OpenGL implementationto use Nvidia vendor library. Relevant for multiple GPUs in a hybrid system (e.g. Intel + Nvidia). Force to use this GPU
# Keep it - Gaming needs this
__GLX_VENDOR_LIBRARY_NAME=nvidia

# Video Acceleration API (VA-API) - HW Acceleration for media players on VLC, Web Browsers, Steam
# Remove if Intel or AMD is wanted for encoding
LIBVA_DRIVER_NAME=nvidia

# Video Decode and Presentation API - enable HW acceleration on video playback (YT, local video files) using NVIDIA instead of Software Decoding
# Remove if Intel or AMD is wanted for decoding
VDPAU_DRIVER=nvidia

seth wrote:

https://wiki.archlinux.org/title/KDE#Pl … e_settings ?

That was it - for the locale issue part, not the video signal loss part.

[motherfucker_big_scorpio@Nexus ~]$ env | grep -E "^LC_|^LANG" | sort
LANG=en_US.UTF-8
LANGUAGE=
LC_ADDRESS=en_AG.UTF-8
LC_COLLATE=en_AG.UTF-8
LC_CTYPE=en_AG.UTF-8
LC_IDENTIFICATION=en_AG.UTF-8
LC_MEASUREMENT=en_AG.UTF-8
LC_MESSAGES=en_AG.UTF-8
LC_MONETARY=pt_PT.UTF-8
LC_NAME=en_AG.UTF-8
LC_NUMERIC=de_DE.UTF-8
LC_PAPER=en_AG.UTF-8
LC_TELEPHONE=en_AG.UTF-8
LC_TIME=de_DE.UTF-8
[motherfucker_big_scorpio@Nexus ~]$

One more note, setting "export LC_ALL=en_US:C" in .bashrc and intending for a functional fallback locale wouldn't work this way as I see this warning:

bash: warning: setlocale: LC_ALL: cannot change locale (en_US:C): No such file or directory

Still, journalctl does not report invalid locale for LC_NUMERIC being en_DE (that it took from "~/.config/plasma-localerc") for steam as we've seen earlier.

---

Many thanks seth and V1del.
This topic has been solved and turned out to be a KDE Plasma Wayland + misconfigured environment variable value issue.

(note from earlier seth: I don't know what to do about drkonqi coredump spamming baloo_file_extractor warnings)

Last edited by Big Scorpio (2026-03-25 20:55:12)

seth · 2026-03-25 21:27:38

\o/
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.

what to do about drkonqi coredump spamming baloo_file_extractor warnings

https://wiki.archlinux.org/title/Core_d … dump_files

Big Scorpio · 2026-03-25 21:52:53

Gotcha.

seth wrote:

Please always remember to mark resolved threads by editing your initial posts subject

I did that before this post of yours. You missed it? Or did not see it?

Last edited by Big Scorpio (Today 22:08:32)

(as of now)

seth · 2026-03-25 21:57:25

"Missed", I had the tab open for longer but got distracted.
Sorry for the noise.

V1del · 2026-03-26 17:11:12

FWIW for the environment question, given that this is the kind of setup you want to use and enforce the ones you've shown there do look like sane choices to coerce the relevant stuff towards nvidia despite the presence of the intel gpu (and then overriding them again for the specific instances you explicitly want intel for)

Arch Linux

#1 2026-03-23 21:11:16

[SOLVED] No video signal after system update & log in - Plasma Wayland

#2 2026-03-23 22:00:16

Re: [SOLVED] No video signal after system update & log in - Plasma Wayland

#3 2026-03-24 19:05:52

Re: [SOLVED] No video signal after system update & log in - Plasma Wayland

#4 2026-03-24 20:35:31

Re: [SOLVED] No video signal after system update & log in - Plasma Wayland

#5 2026-03-24 21:19:51

Re: [SOLVED] No video signal after system update & log in - Plasma Wayland

#6 2026-03-24 21:34:45

Re: [SOLVED] No video signal after system update & log in - Plasma Wayland

#7 2026-03-24 21:36:06

Re: [SOLVED] No video signal after system update & log in - Plasma Wayland

#8 2026-03-25 00:36:12

Re: [SOLVED] No video signal after system update & log in - Plasma Wayland

#9 2026-03-25 01:17:09

Re: [SOLVED] No video signal after system update & log in - Plasma Wayland

#10 2026-03-25 07:50:04

Re: [SOLVED] No video signal after system update & log in - Plasma Wayland

#11 2026-03-25 20:51:36

Re: [SOLVED] No video signal after system update & log in - Plasma Wayland

#12 2026-03-25 21:27:38

Re: [SOLVED] No video signal after system update & log in - Plasma Wayland

#13 2026-03-25 21:52:53

Re: [SOLVED] No video signal after system update & log in - Plasma Wayland

#14 2026-03-25 21:57:25

Re: [SOLVED] No video signal after system update & log in - Plasma Wayland

#15 2026-03-26 17:11:12

Re: [SOLVED] No video signal after system update & log in - Plasma Wayland

Board footer