Long black screen after login and KDE session crash after some minutes

Issues & Assistance
1

Hey everyone,
I am quite new to the OS and still getting my head around it. I have a 7800x3D system with a 4080 card and KDE as DE. At first, everything worked well - apart from the fact I HDMI doesn’t work with my monitor and have to use displayport - but more recently I’ve been experience something that is quite frustrating.

After logging into the system, at least 50% of the times it is taking about a minute for me to get a screen and load the desktop. 100% of the times when this happens, after some minutes, I get back to the login screen automatically. If I try to login again, the system just freezes and I need to hard shutdown.

Investigating a little by myself (with help of ChatGPT/Gemini) I’ve found the following:

Aug 01 21:00:07 Cerberus systemd[949]: Starting KDE Window Manager...

Aug 01 21:00:07 Cerberus systemd[949]: Started KDE Window Manager.

Aug 01 21:00:07 Cerberus kwin_wayland[1019]: No backend specified, automatically choosing drm

Aug 01 21:00:08 Cerberus kwin_wayland[1019]: kf.svg: The theme "Ant-Dark" uses the legacy metadata.desktop. Co>

Aug 01 21:00:08 Cerberus kwin_wayland[1019]: kf.svg: The theme "Ant-Dark" uses the legacy metadata.desktop. Co>

Aug 01 21:00:08 Cerberus kwin_wayland[1019]: kf.plasma.core: The theme "Ant-Dark" uses the legacy metadata.des>

Aug 01 21:00:08 Cerberus kwin_wayland[1019]: kf.plasma.core: The theme "Ant-Dark" uses the legacy metadata.des>

Aug 01 21:00:08 Cerberus kwin_wayland_wrapper[1092]: The XKEYBOARD keymap compiler (xkbcomp) reports:

Aug 01 21:00:08 Cerberus kwin_wayland_wrapper[1092]: > Warning:          Could not resolve keysym XF86RefreshR>

Aug 01 21:00:08 Cerberus kwin_wayland_wrapper[1092]: > Warning:          Could not resolve keysym XF86Accessib>

Aug 01 21:00:08 Cerberus kwin_wayland_wrapper[1092]: > Warning:          Could not resolve keysym XF86DoNotDis>

Aug 01 21:00:08 Cerberus kwin_wayland_wrapper[1092]: Errors from xkbcomp are not fatal to the X server

Aug 01 21:00:08 Cerberus kwin_wayland_wrapper[1100]: The XKEYBOARD keymap compiler (xkbcomp) reports:

Aug 01 21:00:08 Cerberus kwin_wayland_wrapper[1100]: > Warning:          Unsupported maximum keycode 708, clip>

Aug 01 21:00:08 Cerberus kwin_wayland_wrapper[1100]: >                   X11 cannot support keycodes above 255.

Aug 01 21:00:08 Cerberus kwin_wayland_wrapper[1100]: > Warning:          Could not resolve keysym XF86RefreshR>

Aug 01 21:00:08 Cerberus kwin_wayland_wrapper[1100]: > Warning:          Could not resolve keysym XF86Accessib>

Aug 01 21:00:08 Cerberus kwin_wayland_wrapper[1100]: > Warning:          Could not resolve keysym XF86DoNotDis>

Aug 01 21:00:08 Cerberus kwin_wayland_wrapper[1100]: Errors from xkbcomp are not fatal to the X server

Aug 01 21:00:08 Cerberus kcminit[1101]: Initializing  "/usr/lib/qt6/plugins/plasma/kcms/systemsettings/kcm_fon>

Aug 01 21:00:08 Cerberus kcminit[1101]: Initializing  "/usr/lib/qt6/plugins/plasma/kcms/systemsettings/kcm_sty>

Aug 01 21:00:10 Cerberus kwin_wayland[1019]: Module 'org.kde.kwin.decoration' does not contain a module identi>

Aug 01 21:00:10 Cerberus kwin_wayland[1019]: kf.svg: The theme "Ant-Dark" uses the legacy metadata.desktop. Co>

Aug 01 21:00:13 Cerberus kwin_wayland[1019]: kf.windowsystem: static bool KX11Extras::mapViewport() may only b>

Aug 01 21:00:15 Cerberus kwin_wayland[1019]: kf.windowsystem: static bool KX11Extras::mapViewport() may only b>

Aug 01 21:01:07 Cerberus kwin_wayland[1019]: kwin_wayland_drm: atomic commit failed: Permission denied

Aug 01 21:01:17 Cerberus kwin_wayland[1019]: kwin_screencast: PipeWire remote error:  connection error

Aug 01 21:01:22 Cerberus systemd[949]: Stopping KDE Window Manager...

Aug 01 21:01:22 Cerberus systemd[949]: Stopped KDE Window Manager.

Aug 01 21:01:22 Cerberus systemd[949]: plasma-kwin_wayland.service: Consumed 11.818s CPU time, 164M memory

This is the log straight after the crash. The issue seems to be here:

Aug 01 21:01:07 Cerberus kwin_wayland[1019]: kwin_wayland_drm: atomic commit failed: Permission denied

And then

Aug 01 21:01:17 Cerberus kwin_wayland[1019]: kwin_screencast: PipeWire remote error:  connection error
Aug 01 21:01:22 Cerberus systemd[949]: Stopping KDE Window Manager...
Aug 01 21:01:22 Cerberus systemd[949]: Stopped KDE Window Manager.

Has anyone seen this issue before? Is this a known issue for Nvidia?

I would love to get some assistance to fix it. Important to note it does not happen 100% of the times. If I get a long time to get into the DE, which makes me know that I will have the crash I just manually reboot and start a new session until I get the normal behavior (instant DE being displayed after login).

I am not currently at the system but I will update with Kernel and driver version, but I keep it up to date so assume everything is the latest.

Thanks!

2

System information can be provided via;

sudo cachyos-bugreport.sh

Do you happen to have some sort of splash screen applied?
System Settings > Appearance & Style > Colors & Themes > Splash Screen

This sounds like a systemic problem.

Your theme is outdated.
This could be causing any number of issues.
Please choose something else - the standard is Breeze .. it may be worthwhile to apply defaults/breeze in the various System Settings sections. The “Global” themes do this in a bundled way which may or may not be desirable.

More theme problems, including for window decorations.

Possibly permissions problems, but not necessarily.

Please go through the earlier sections of this message and report back.

3

Thank you my friend.
I had no splash screen set. Reverted back to breeze, I’ll report back if the issue happens again!

4

I was not suggesting to apply a splash.
( Here splash screen “None” is used and it does not cause a long delay. )
I was guessing the issue might be related to splash .. possibly a broken theme.
If a problem is created only when using None for splash then it is maybe a little more mysterious.

5

Hey @cscs yep I got what you meant not suggesting to apply a splash screen!
I think you were on point with the outdated theme. Since I reverted back to Breeze I got no issues. Will keep monitoring and report if it happens again.

6

Hey @cscs , the issue happened again. I am not using any custom theme, I have the general Breeze Dark selected.

I’ve generated the log in here.

Reviewed everything theme related, I have everything set to default (Breeze Dark), through global theme and then manually checking each option.

7

I also notice a bunch of USB failures?

Do you have some peripherals plugged in?

Otherwise I might also ask for something like;

systemd-analyze blame

systemd-analyze critical-chain  

systemd-analyze plot > boot_plot.svg

(The last one will produce an svg chart of the boot.)

8

Hi,

My friend and I have had the same experience(s) before.
I don’t remember if it always after an update, but some times, as you said, one logins in, after few seconds gets logged out, enters password again, stays at login screen and complete freeze.
Hard restart required.
I don’t know what causes this, perhaps it is not cachyos-related, but of Arch/KDE?

Best,

9

Hello,

I have a similar issue with a fully patched freashly installed CachyOS as of 8-31-25 on both the stable kernal, 6.14, and LTS kernal, 6.12, WITH some unqiue differences below. I am using the default KDE theme and SDDM settings out of the box/install process. I installed the OS two weeks ago with GRUB as the boot loader.

If I login into KDE using Wayland or X11, the log process/splash screen is slow to show up. It will be a blank screen first, then the slash screen process show up 5 secs later and then it completes the login process and the desktop shows up.

My unique issue is that if I log into KDE X11 from boot FIRST, I get the same issue as above AND after the desktop fully loads, 20 to 30 secs later, my Nvida dual screens goes blank/black. The OS is not responsive and I have to power down. The TTY terminal session options are not responsive either.

However, if I boot up CachyOS and login into Wayland FIRST, log out when everything loads, then login into KDE using X11, the above unique issue (where X11 goes to a blank screen after 20-30 secs after the desktop load and the OS become unrespondsive/crashes) does NOT occurr and X11 is fine until the next reboot.

Wayland is not stable with Firefox or Brave like other Arch distro where its UI windows rendering is off, the browser flashes in out, or freezes when you move or resize the browser windows between the dual screens or within the same monitor, so I use X11.

And I have few/rare experences in Wayland only where after logging in, it will automatically log out after the desktop fully loads within a 20 secs as others in this post talked about also.

Help! It seem like there are universal issues with CachyOS – a blank unresposive OS black screen after X11 primary logins, slow KDE login splash screen load up, and on rare occassions, the OS auto log outs 20 secs into use in Wayland.

Other fully patched Arch KDE rolling release distro does not have these special CachyOS “features” as dicussed on this forum using the SAME hardware. I really really want to give CachyOS a try.

Please fix CachyOS, if possible. TY.

10

I also have the same problem. It may occur several times in a row or it may only occur once. I did notice some casual correlations: if I had no applications open from the last session before reboot it was much less likely to occur but still does occur anyway occasionally; if I see the C (for Cachy) boot splash fully it is less likely to occur but still does occur anyway occasionally; if the black screen between login and desktop loading is longer than 5 seconds it is much more likely to occur with a longer duration significantly increasing the likelihood; and when I am returned automatically by this bug from the desktop session to the login screen I only have about 5 seconds to hit reboot before the system freezes completely requiring a hard reboot but if I do hit reboot in time I am less likely to have a reoccurrence of the issue as compared to after going through a hard reboot although it does happen in both scenarios - sometimes requiring a reboot/hard boot several times to auto correct itself. Once I have been successfully logged into a Wayland session for more than 5 minutes it has never occured.

11

Maybe worth trying out. Currently i do not have the blackscreen and freezes any longer. (Fingers crossed it stays this way :D)

12

Thanks Takinya for the advise. I did try several options but all failed for me at least. I tried disabling my iGPU in the BIOS settings only, only setting Cachyos/KDE to always start a new session after restarting and not try to restore an old one, and trying both options simultaneously.

13

Mhh do you have your sudo cachyos-bugreport.sh logs?

14

It took me several boots to replicate the issue, as it was acting much better for some reason recently, as it has been boot up fine about 80% of the time. Now I even have a video of the session in which the bug.log report was taken where it abruptly reloads the login screen from which I chose to reboot from. I didn’t have any problem with the resulting Wayland session after the reboot at least. That is, unfortunately, not always the case but at least the significantly long time lag between the first login and the Wayland session is so predictable of forthcoming failure that I was able to be prepared and capture everything in advance as it happened.

15

[ 0.410131] pci_bus 0000:00: Some PCI device resources are unassigned, try booting with pci=realloc

It looks like a GPU-problem too, i would say :smiley:

16

It’s strange that this is an intermittent issue for you. You said that all was well and then you started having this issue which makes it seem as if this happened out of the blue BUT something must have changed beforehand. Did you update any packages related to your graphics card or change/update the driver? From the terminal run

inxi -Fza

for a complete list/output of your hardware both internal and external (the flags are F for full, z for security filters, a for admin level output/more details) and the drivers being used. Or just do

inxi -Gza

to see only info related to your graphics card. Additionally, PipeWire is a multimedia server used for things like screen sharing/recording and its failure seems to be directly related to what appears to be intermittent graphics card issues that are possibly driver related. Also, if you haven’t already, disable the 7800’s integrated graphics in your bios just to be sure there isn’t some sort of conflict going on.

17

Here is the output for: ❯ inxi -Gza
Graphics:
Device-1: NVIDIA GB202 [GeForce RTX 5090] vendor: ASUSTeK driver: nvidia
v: 590.48.01 alternate: nouveau,nvidia_drm non-free: 550-580.xx+
status: current (as of 2025-11) arch: Lovelace code: AD1xx
process: TSMC n4 (5nm) built: 2022+ pcie: gen: 5 speed: 32 GT/s lanes: 16
ports: active: HDMI-A-1 empty: DP-1, DP-2, DP-3, HDMI-A-2 bus-ID: 01:00.0
chip-ID: 10de:2b85 class-ID: 0300
Display: wayland server: X.org v: 1.21.1.21 with: Xwayland v: 24.1.9
compositor: kwin_wayland driver: gpu: nv_platform,nvidia,nvidia-nvswitch
display-ID: 0
Monitor-1: HDMI-A-1 model: Samsung QBQ90 serial: built: 2022 res:
mode: 1920x1080 hz: 60 scale: 100% (1) dpi: 52 gamma: 1.2
size: 1872x1053mm (73.7x41.46") diag: 2148mm (84.6") ratio: 16:9 modes:
max: 3840x2160 min: 640x480
API: EGL v: 1.5 hw: drv: nvidia platforms: device: 0 drv: nvidia device: 2
drv: swrast gbm: drv: nvidia surfaceless: drv: nvidia wayland: drv: nvidia
x11: drv: nvidia inactive: device-1
API: OpenGL v: 4.6.0 compat-v: 4.5 vendor: nvidia mesa v: 25.3.2-arch1.2
glx-v: 1.4 direct-render: yes renderer: NVIDIA GeForce RTX 5090/PCIe/SSE2
memory: 31.1 GiB display-ID: :0.0
API: Vulkan v: 1.4.335 layers: 12 device: 0 type: discrete-gpu
name: NVIDIA GeForce RTX 5090 driver: nvidia v: 590.48.01
device-ID: 10de:2b85 surfaces: N/A
Info: Tools: api: clinfo, eglinfo, glxinfo, vulkaninfo
de: kscreen-console,kscreen-doctor gpu: nvidia-settings,nvidia-smi
wl: wayland-info x11: xdpyinfo, xprop, xrandr

~
❯

I did change the settings in the BIOS to disable the built in Graphics Card. I think I did it before but I will give it a try again just in case I reset everything before giving it a go. I will let you know how it went in a few weeks. To recap: the issue appears randomly, sometimes once and sometimes over multiple consecutive reboots, yet the screen freeze is always predictable based on on the time delay between entering the login screen credentials then hitting and the first visual confirmation of the desktop session loading. If the desktop session visually loads in less than 5 seconds, I am good to go 100% of the time. If the black screen lag time, between login screen credential and the first visual confirmation of the desktop session, is longer than about 5 seconds (which most of the time in this situation it is significantly longer such as around 15-45 seconds) it will fail 100% of the time after showing the desktop session for about 3-5 minutes then spontaneously loading the login screen and freezing after 5 seconds if I don’t reboot the system before then.