Sudden Blackscreen during normal use in Plasma DE

Arlunlux · 2024-12-23 17:11:33

Hey,

for quite some time, I encounter a strange problem with my Arch installation using the KDE Plasma Desktop Environment. The problem last occurred today, on a fresh installation (but same hardware), but I experienced the problem two times before that, during updates, which killed this installation. They might also be unrelated, but I don't think so.

I would really appreciate learning about two things:
- does anyone have an explanation for this behavior? What are good resources to debug/analyze this?
- what are good recovery steps that I can try next time before doing a hard reset? Are there some magic recovery keys (you can set up)?

EDIT: I now know about sysRQ, I will try it out next time something like this happens.

Thanks in advance,
Luna

Description

I was just doing some light browsing on firefox, logged into my normal user (is in wheel group), staring at a page. Without any direct action or input, the screen turned black, but the LEDs on the keyboard and power button seemed unaffected. For reference in the logs, the time was 15:40 (3:40 pm) when I looked at my phone to check the time of the failure. Logging into a different TTY didn't work (Ctrl + Alt + F2/3/4), so I hard reset the Laptop. Except for some orphan inodes, I think I got away this time.

System Information

Kernel: linux 6.12.6.arch1-1
KDE Plasma: 6.2.4
KDE Frameworks: 6.9.0
Qt: 6.8.1
Wayland

Hardwareinfo (dmidecode):

Getting SMBIOS data from sysfs.
SMBIOS 3.3.0 present.

Handle 0x0000, DMI type 0, 26 bytes
BIOS Information
Vendor: LENOVO
Version: JPCN20WW(V1.06)
Release Date: 09/20/2022
Address: 0xE0000
Runtime Size: 128 kB
ROM Size: 16 MB
Characteristics:
PCI is supported
BIOS is upgradeable
BIOS shadowing is allowed
Boot from CD is supported
Selectable boot is supported
EDD is supported
Japanese floppy for NEC 9800 1.2 MB is supported (int 13h)
Japanese floppy for Toshiba 1.2 MB is supported (int 13h)
5.25"/360 kB floppy services are supported (int 13h)
5.25"/1.2 MB floppy services are supported (int 13h)
3.5"/720 kB floppy services are supported (int 13h)
3.5"/2.88 MB floppy services are supported (int 13h)
8042 keyboard services are supported (int 9h)
CGA/mono video services are supported (int 10h)
ACPI is supported
USB legacy is supported
BIOS boot specification is supported
Targeted content distribution is supported
UEFI is supported
BIOS Revision: 1.20
Firmware Revision: 1.20

Handle 0x0001, DMI type 1, 27 bytes
System Information
Manufacturer: LENOVO
Product Name: 21DL
Version: ThinkBook 15 G4 ABA
Serial Number: omitted
UUID: omitted
Wake-up Type: Power Switch
SKU Number: omitted
Family: ThinkBook 15 G4 ABA

Logs

Journal starting 10 min before the black screen and ending after the initial bootup after the hard reset: https://0x0.st/8rjn.txt
New log, freeze after suspend: https://0x0.st/s/w_eViEpb1dHT2m7T0Qv0IQ/8r6g.txt
(you can watch me frantically trying to use sysRQ and a different keyboard haha)

Installed packages:
amd-ucode 20241210.b00a7f7e-1, base 3-2, bluedevil 1:6.2.4-1, breeze 6.2.4-1, breeze-gtk 6.2.4-1, breeze-plymouth 6.2.4-1, discover 6.2.4-1, dolphin 24.12.0.1-1, dolphin-plugins 24.12.0-1, drkonqi 6.2.4-1, firefox 133.0.3-2, flatpak-kcm 6.2.4-1, fuse2 2.9.9-5, hyphen-de 20220923-2, hyphen-en 2.8.8-6, isoimagewriter 24.12.0-1, iwd 3.3-1, kactivitymanagerd 6.2.4-1, kate 24.12.0-1, kde-cli-tools 6.2.4-1amd-ucode 20241210.b00a7f7e-1, base 3-2, bluedevil 1:6.2.4-1, breeze 6.2.4-1, breeze-gtk 6.2.4-1, breeze-plymouth 6.2.4-1, discover 6.2.4-1, dolphin 24.12.0.1-1, dolphin-plugins 24.12.0-1, drkonqi 6.2.4-1, firefox 133.0.3-2, flatpak-kcm 6.2.4-1, fuse2 2.9.9-5, hyphen-de 20220923-2, hyphen-en 2.8.8-6, isoimagewriter 24.12.0-1, iwd 3.3-1, kactivitymanagerd 6.2.4-1, kate 24.12.0-1, kde-cli-tools 6.2.4-1, kde-gtk-config 6.2.4-1, kdecoration 6.2.4-1, kdeplasma-addons 6.2.4-1, keepassxc 2.7.9-4, kgamma 6.2.4-1, kglobalacceld 6.2.4-1, kinfocenter 6.2.4-1, kmenuedit 6.2.4-1, konsole 24.12.0-1, kpipewire 6.2.4-1, krdp 6.2.4-1, kscreen 6.2.4-1, kscreenlocker 6.2.4-1, ksshaskpass 6.2.4-1, ksystemlog 24.12.0-1, ksystemstats 6.2.4-1, kwallet-pam 6.2.4-1, kwayland 6.2.4-1, kwin 6.2.4-2, kwrited 6.2.4-1, layer-shell-qt 6.2.4-2, libkscreen 6.2.4-1, libksysguard 6.2.4-1, libmythes 1:1.2.5-2, libplasma 6.2.4-1, libreoffice-still 24.2.7-3, linux 6.12.6.arch1-1, linux-firmware 20241210.b00a7f7e-1, linux-firmware-qlogic 20241210.b00a7f7e-1, lvm2 2.03.29-1, man-db 2.13.0-1, milou 6.2.4-1, mythes-de 20240703-1, mythes-en 20240519-1, nano 8.2-1, ocean-sound-theme 6.2.4-1, okular 24.12.0-1, opendoas 6.8.2-3, oxygen 6.2.4-1, oxygen-sounds 6.2.4-1, packagekit-qt5 1.1.2-1, plasma-activities 6.2.4-1, plasma-activities-stats 6.2.4-1, plasma-browser-integration 6.2.4-1, plasma-desktop 6.2.4-1, plasma-disks 6.2.4-1, plasma-firewall 6.2.4-1, plasma-integration 6.2.4-1, plasma-nm 6.2.4-1, plasma-pa 6.2.4-1, plasma-sdk 6.2.4-1, plasma-systemmonitor 6.2.4-1, plasma-thunderbolt 6.2.4-1, plasma-vault 6.2.4-1, plasma-welcome 6.2.4-1, plasma-workspace 6.2.4-1, plasma-workspace-wallpapers 6.2.4-1, plasma5support 6.2.4-1, plymouth-kcm 6.2.4-1, polkit-kde-agent 6.2.4-1, power-profiles-daemon 0.23-2, powerdevil 6.2.4-1, print-manager 1:6.2.4-1, qqc2-breeze-style 6.2.4-1, sddm-kcm 6.2.4-1, spotify-launcher 0.6.0-1, sweeper 24.12.0-1, systemsettings 6.2.4-1, ttf-liberation 2.1.5-2, wacomtablet 6.2.4-1, xdg-desktop-portal-kde 6.2.4-1, zsh 5.9-5, , kde-gtk-config 6.2.4-1, kdecoration 6.2.4-1, kdeplasma-addons 6.2.4-1, keepassxc 2.7.9-4, kgamma 6.2.4-1, kglobalacceld 6.2.4-1, kinfocenter 6.2.4-1, kmenuedit 6.2.4-1, konsole 24.12.0-1, kpipewire 6.2.4-1, krdp 6.2.4-1, kscreen 6.2.4-1, kscreenlocker 6.2.4-1, ksshaskpass 6.2.4-1, ksystemlog 24.12.0-1, ksystemstats 6.2.4-1, kwallet-pam 6.2.4-1, kwayland 6.2.4-1, kwin 6.2.4-2, kwrited 6.2.4-1, layer-shell-qt 6.2.4-2, libkscreen 6.2.4-1, libksysguard 6.2.4-1, libmythes 1:1.2.5-2, libplasma 6.2.4-1, libreoffice-still 24.2.7-3, linux 6.12.6.arch1-1, linux-firmware 20241210.b00a7f7e-1, linux-firmware-qlogic 20241210.b00a7f7e-1, lvm2 2.03.29-1, man-db 2.13.0-1, milou 6.2.4-1, mythes-de 20240703-1, mythes-en 20240519-1, nano 8.2-1, ocean-sound-theme 6.2.4-1, okular 24.12.0-1, opendoas 6.8.2-3, oxygen 6.2.4-1, oxygen-sounds 6.2.4-1, packagekit-qt5 1.1.2-1, plasma-activities 6.2.4-1, plasma-activities-stats 6.2.4-1, plasma-browser-integration 6.2.4-1, plasma-desktop 6.2.4-1, plasma-disks 6.2.4-1, plasma-firewall 6.2.4-1, plasma-integration 6.2.4-1, plasma-nm 6.2.4-1, plasma-pa 6.2.4-1, plasma-sdk 6.2.4-1, plasma-systemmonitor 6.2.4-1, plasma-thunderbolt 6.2.4-1, plasma-vault 6.2.4-1, plasma-welcome 6.2.4-1, plasma-workspace 6.2.4-1, plasma-workspace-wallpapers 6.2.4-1, plasma5support 6.2.4-1, plymouth-kcm 6.2.4-1, polkit-kde-agent 6.2.4-1, power-profiles-daemon 0.23-2, powerdevil 6.2.4-1, print-manager 1:6.2.4-1, qqc2-breeze-style 6.2.4-1, sddm-kcm 6.2.4-1, spotify-launcher 0.6.0-1, sweeper 24.12.0-1, systemsettings 6.2.4-1, ttf-liberation 2.1.5-2, wacomtablet 6.2.4-1, xdg-desktop-portal-kde 6.2.4-1, zsh 5.9-5

Last edited by Arlunlux (2024-12-25 12:24:37)

V1del · 2024-12-24 13:08:22

Nothing obvious in the journal excerpt. Reproduce this and use sysrq then post the resultant journal.

FWIW are you potentially running into the auto suspend timeout? There are some reports with crashes when trying to wake up from this especially in correlation with the mediatek chip that seems to be used for wifi here.

on second look some opengl errros potentially with allocating new resources. And apparently just 2GB VRAM, can you check your UEFI on whether it's possible to allocate more system memory to VRAM?

FWIW there's also https://bbs.archlinux.org/viewtopic.php?id=301798&p=2 suggesting a regression with certain APUs and mesa 24.3 maybe try downgrading mesa to 24.2.7

Last edited by V1del (2024-12-24 13:51:27)

Arlunlux · 2024-12-24 14:14:20

V1del wrote:

Nothing obvious in the journal excerpt. Reproduce this and use sysrq then post the resultant journal.
FWIW are you potentially running into the auto suspend timeout? There are some reports with crashes when trying to wake up from this especially in correlation with the mediatek chip that seems to be used for wifi here.

Actually, now that you say that, I actually had problems in the past with the system not waking up from a suspend, especially after leaving it in suspend for a while, even with a different Linux distro installed. Might be worth looking into.
EDIT: a (normal) suspend is unlikely, since the power LED goes into a pulsing mode when suspended. This was definitely not the case with the last crash, it stayed solid white. I think I just have to wait and see until the next time this happens.

Tbh, I hate how laptops always have to have some obscure chips with bad firmware and drivers on/for them.

Last edited by Arlunlux (2024-12-24 14:29:19)

corpest · 2024-12-24 17:52:51

"I had something similar happening, I took the lazy approach, and just switched to GNOME, and that seemed to fix it.

rg_arc · 2024-12-24 18:02:34

I wonder if it has anything to do with wayland because in sway (I know its not at all like Plasma DE) there was a bug like this: https://github.com/swaywm/sway/issues/8500 and there were several bugs similar to that I've seen not just with sway.
I can confirm that issue but I used the git version of sway and wlroots and that solved the problem (at least for now). It may not fix your issue but perhaps try to see if it has something to do with wayland?

Last edited by rg_arc (2024-12-24 18:04:11)

seth · 2024-12-24 19:03:26

This is a GFX9 chip, see https://bbs.archlinux.org/viewtopic.php … 6#p2216126
Switching to gnome will do jackshit about that.

Arlunlux · 2024-12-24 20:28:42

for clarification: I was just looking at the screen for a few seconds, I find it very unlikely that this is a suspend issue.

Arlunlux · 2024-12-25 10:44:03

Wow, I accidentally left the system at idle and it suspended over night, I let it wake up rn and now I logged in normally, but then, shortly after, the system went unresponsive (the windows greyed out, even rendering (unresponsive). The magic sysRQ keys don't work (all of them), but weirdly, I can still open new terminals with Ctrl+Alt+T and move the mouse around freely. At the start, switching to a virtual console worked (ctrl+Alt+F3), but no keyboard input.

I am completely confused, I can still open as many terminals as I like with the key combination, but no typing/interaction by clicking.

EDIT: Even Alt-Tab works, so some process still seems to work. And I'm sure I press the correct sysRQ because I tested the combination with evtest. So how could Plasma possibly refuse SysRQ+r and SysRQ+k?

EDIT 2: Okay, I apparently messed up the sysrq setup, it was set to kernel.sysrq = 16 (systemd default?). Going to post the new logs later today. (added them to the inital post)

Last edited by Arlunlux (2024-12-25 12:25:10)

seth · 2024-12-25 14:45:14

The magic sysRQ keys don't work (all of them), but weirdly, I can still open new terminals with Ctrl+Alt+T

The sysrq feature has to be explicitly enabled (see the wiki) - there's essentially no way that it failed if the userspace still responded.

Okay, I apparently messed up the sysrq setup

… aaaand I should really first read the entire post before responding

The recent condition is however likely different from

Logging into a different TTY didn't work (Ctrl + Alt + F2/3/4), so I hard reset the Laptop.

the windows greyed out, even rendering (unresponsive)

Did you btw. check the link in #6 and try to set "amdgpu.mcbp=0" and/or downgrade mesa as V1del suggested?

Arlunlux · 2025-02-09 17:10:40

Hey, it has been a while, but the issue has gotten much worse. It sometimes takes multiple attempts to boot my system successfully. On some occasions, I just get a blackscreen after decrypting my drive, sometimes the computer completely shuts off. Yesterday, my laptop just went out during normal browsing. The windows install on my second SSD seems unaffected, live boot disks too. I also seem to get some warnings during early boot and when shutting down, but I don't know how to access these logs. After my last post on this thread, I installed hyprland alongside Plasma, the packet list is the below. My current journal since this successfull boot:

https://0x0.st/s/7Z_a1WR5tILWoNKqVIpEVQ/8PnP.txt

and no, I didn't set

amdgpu.mcbp=0

I didn't really understand the function of that.

Package List:
https://0x0.st/s/qAyQyJAXBXz8SoMcoJ7Jew/8ZPI.txt

Last edited by Arlunlux (2025-02-10 20:30:27)

seth · 2025-02-09 22:03:32

Please use [code][/code] tags. Edit your post in this regard.

completely shuts off. Yesterday, my laptop just went out during normal browsing. The windows install on my second SSD seems unaffected

3rd link below. Mandatory.
Disable it (it's NOT the BIOS setting!) and reboot windows and linux twice for voodo reasons.

There's is an enormous chance that windows is "shutting off" your PC (via an acpi schedule entry) to do its matainanace stuff.
This would also explain the live-system behavior, since the firmware understands that you're not resuming a hibernated system here.

Arlunlux · 2025-02-10 20:47:44

seth wrote:

Please use [code][/code] tags. Edit your post in this regard.

I edited accordingly and also put the packet list on a pastebin for more readability.

3rd link below. Mandatory.
Disable it (it's NOT the BIOS setting!) and reboot windows and linux twice for voodo reasons.

Did that, no change

There's is an enormous chance that windows is "shutting off" your PC (via an acpi schedule entry) to do its matainanace stuff.

Is there a way to dump the acpi context in a livedisk?

Btw, multi-user.target for systemd also doesnt work. Only emergency mode works, rescue doesn't.

My next step is installing acpica via the livedisk to dump the acpi context in emergency mode

EDIT: now the livedisk also doesn't work. Trying to dump the acpi context via windows now...

EDIT: wow, even windows blacked out on me during startup. Beginning to think this might be a deeper hardware/firmware issue. I just plugged it in and now Linux works, but I definitely had times where Linux didn't start even with the laptop being plugged in. My next laptop will 100% something with less OEM trash/buggyness.

Last edited by Arlunlux (2025-02-10 20:59:33)

Arlunlux · 2025-02-10 21:08:27

Output of:

acpidump -b && for i in *; do echo $i:; strings -a $i | grep -i windows; done

can be found here: https://0x0.st/s/ZQjRMN2h_to7TTBRehL2Hw/8ZZa.log

EDIT: I searched for new bios updates and there was one from Nov 24 stating this in the readme:
Update the AMD GOP Driver to resolve the POST black screen issue on some screens

could this be related?

Last edited by Arlunlux (2025-02-10 21:18:10)

seth · 2025-02-10 22:50:50

If it's not the GPU and wrt the sudden shutdowns and assuming that's also not windows but rather affects it, too:
https://wiki.archlinux.org/title/Ryzen#Troubleshooting

Ryzens are notorious for that and it's typically an aggravating problem as well.
Try to increase the voltage to the CPU (if the firmware allows that) and/or limit teh c-states (as it's often triggered by core cycling)

Arlunlux · 2025-02-10 23:02:45

I'll test the new UEFI Update for the next few days and keep this thread updated if the black screens happen again. I deeply hope it is not a C-State or Voltage Problem

Arch Linux

#1 2024-12-23 17:11:33