I just reinstalled my system because I wanted a clean slate for the upcoming Plasma 6. System was running fine otherwise, although there was a ton of Gnome software from switching DEs. I can’t put system to sleep/suspend, it immediately wakes up, fans turn on, screen goes black and won’t respond to any keyboard inputs. Have to cold boot by removing PSU cable, case power button won’t shut down system. It’s an RX 7800 XT GPU.
Any ideas?
Bug similar to this: 7800XT amdgpu initialisation fails on reboot with kernel 6.7 (#3062) · Issues · drm / amd · GitLab
Nothing about AMDGPU on log.
sudo journalctl -o short-precise -k -b -1 -r -p 4
Feb 26 14:17:51.603166 klaasjew kernel: warning: `kded5' uses wireless extensions which will stop working for Wi-Fi 7 hardware; use nl80211
Feb 26 14:17:49.931697 klaasjew systemd-journald[986]: File /var/log/journal/b0c27d05b32647c29d25eb4cca1d3150/user-1000.journal corrupted or uncleanly shut down, renaming and replacing.
Feb 26 14:17:37.519697 localhost.localdomain kernel: block nvme1n1: No UUID available providing old NGUID
Feb 26 14:17:36.539720 localhost.localdomain kernel: Bluetooth: hci0: Malformed MSFT vendor event: 0x02
Feb 26 14:17:34.969915 localhost.localdomain kernel: iwlwifi 0000:06:00.0: WRT: Invalid buffer destination
Feb 26 14:17:34.669713 localhost.localdomain kernel: iwlwifi 0000:06:00.0: WRT: Invalid buffer destination
Feb 26 14:17:33.589696 localhost kernel: iwlwifi 0000:06:00.0: WRT: Invalid buffer destination
Feb 26 14:17:33.583182 localhost kernel: thermal thermal_zone0: failed to read out thermal zone (-61)
Feb 26 14:17:33.505952 localhost kernel: Bluetooth: hci0: No support for _PRR ACPI method
Feb 26 14:17:33.505394 localhost kernel: iwlwifi 0000:06:00.0: api flags index 2 larger than supported by driver
Feb 26 14:17:33.504812 localhost systemd-journald[986]: File /var/log/journal/b0c27d05b32647c29d25eb4cca1d3150/system.journal corrupted or uncleanly shut down, renaming and replacing.
Feb 26 14:17:33.356367 localhost kernel: ACPI Warning: SystemIO range 0x0000000000000B00-0x0000000000000B08 conflicts with OpRegion 0x0000000000000B00-0x0000000000000B0F (\GSA1.SMBI) (20230628/utaddress-204)
Feb 26 14:17:32.793559 localhost systemd[1]: /usr/lib/systemd/system/plymouth-start.service:15: Unit uses KillMode=none. This is unsafe, as it disables systemd's process lifecycle management for the service. Please update the service to use a safer KillMode=, such as 'mixed' or 'control-group'. Support for KillMode=none is deprecated and will eventually be removed.
Feb 26 14:17:27.153329 localhost kernel: ata1.00: supports DRM functions and may not be fully accessible
Feb 26 14:17:27.150027 localhost kernel: ata1.00: supports DRM functions and may not be fully accessible
Feb 26 14:17:25.770947 localhost systemd[1]: /usr/lib/systemd/system/plymouth-start.service:15: Unit uses KillMode=none. This is unsafe, as it disables systemd's process lifecycle management for the service. Please update the service to use a safer KillMode=, such as 'mixed' or 'control-group'. Support for KillMode=none is deprecated and will eventually be removed.
Enable magic SysRq key combo from Yast so you can at least attempt a safe reboot:
The above may not help as it sounds like you’re having a kernel panic.
You would need to setup kdump to figure out what’s going on:
Thank you for the information and links. Unfortunately /var/crash is empty.
This is so frustrating.
It’s empty even after you triggered a crash as per the docs? 
It could happen for example if you have encryption enabled for the volume where /var/crash resides. The kdump kernel needs to mount /var/crash without any password or other user input when dumping.
Yeah, that makes total sense, all my volumes are encrypted. Is there a way for kdump to decrypt the save path?
Also I tried installing again and the bug is there, trigger sleep, system goes and then it wakes up immediately and then hangs.
After that failure I installed Fedora KDE and the exact same thing happened on the same, just formatted disk. It has to be a kernel regression and the journal keeps showing iwlwifi firmware crashes, except when triggering sleep.
Ah, in this case you would have to dump to an external USB formatted with ext4 and mounted at /var/crash (entry should be available in /etc/fstab, make sure you it can fail gracefully being a USB) or shrink the current encrypted volume (shrink order FS → LV → encrypted PV) so it can make room for a new ext4 partition that can be mounted at /var/crash.
I only got the system to write an empty lost-found directory in /var/crash…
I’m confused by the Kexec ad Kdump documentation. Do I have to run kexec via CLI AND ALSO the Yast-Kdump module? Then after restarting, do I load the current kernel or the debug kernel? Or does Yast-Kdump configure the system to load the debug kernel I specified?
That’s created by when fsck runs on the ext4 FS.
Regarding the debug kernel, you don’t need it. Yes! I was surprised by this too, for example in Debian I had to pass in the debug kernel for taking kdump but Suse doesn’t need that 
. Just the kernel-default will do in most cases as the kdump config would turn off most modules that might cause a crash, speaking of which my earlier suggestion to use a USB for the dump might need some reconfiguration of kdump to allow the usbcore module.
Here’s my personal notes for configuring kexec and kdump. I use kexec in lieu of reboot as my firmware takes a long time to load making the normal reboot process very slow.
Setup kexec and kdump:
zypper in kexec-tools makedumpfile kdump yast2-kdump crash kdumpid
systemctl enable --now kexec-load.service
Use kexec for reboot:
ln -s /usr/lib/systemd/system/kexec.target /etc/systemd/system/reboot.target
Configure kdump:
yast2 kdump
systemctl reboot
Test taking kdump:
Make sure kdump.service is active. Logout of DE. Switch to console tty and login as root:
sync && echo c > /proc/sysrq-trigger
On fresh boot:
cd /var/crash/2024-01-29-17-48
cat dmesg
kdumpid vmcore
To analyze the vmcore:
zypper repos
zypper mr --enable repo-debug
zypper refresh
zypper install kernel-default-debuginfo
crash /usr/lib/modules/6.7.1-2-default/vmlinux.xz vmcore
kexec is giving me trouble:
Loaded: loaded (/usr/lib/systemd/system/kexec-load.service; enabled; preset: disabled)
Active: failed (Result: exit-code) since Wed 2024-02-28 19:07:57 CST; 24s ago
Docs: man:kexec(8)
Process: 476 ExecStart=/usr/sbin/kexec-bootloader (code=exited, status=1/FAILURE)
Main PID: 476 (code=exited, status=1/FAILURE)
@lavadrop The question is, why is it enabled? Default is disabled…
Your guess is as good as mine.
@lavadrop @pavinjoseph should be able to provide more information on that as it was their suggestion 
1 Like
Run sudo kexec-bootloader --debug for details.
The question is, why is it enabled? Default is disabled…
This is required for systemctl kexec to work, also generally to test if kexec is working correctly.
It’s looking for the wrong kernel, which I already removed.
Image : /boot/vmlinuz-6.7.6-1-debug
Initrd : /boot/initrd-6.7.6-1-debug
Append : splash=silent resume=/dev/system/swap mitigations=auto quiet security=apparmor
Root : root=/dev/mapper/system-root
Kexec call: kexec --kexec-syscall-auto --load '/boot/vmlinuz-6.7.6-1-debug' --initrd='/boot/initrd-6.7.6-1-debug' --append='root=/dev/mapper/system-root splash=silent resume=/dev/system/swap mitigations=auto quiet security=apparmor'
Cannot open `/boot/vmlinuz-6.7.6-1-debug': No such file or directory
kexec failed.
The default kernel can be seen with:
sudo pbl --default-settings
If it’s sticking to the old kernel, force reinstall the new one:
sudo zypper in --force kernel-default
This command appears to require elevated privileges, does it not?
Yep, everything from the personal notes I shared requires to be run as root.
It’s from my setup notes so being run as root was implicit, sorry for not specifying it.
cli yast2 won’t let me activate kdump, is it the same if I use the GUI version?
Hmm, that’s strange. The TUI and GUI are interfaces to the same thing but who knows, it’s worth a try to check in the GUI too 
What’s the output of:
zypper se -i kexec-tools makedumpfile kdump yast2-kdump crash kdumpid
sudo systemctl status kexec-load.service
S | Name | Summary | Type
---+-------------------+----------------------------------------------------------------------------+--------
i+ | crash | Crash utility for live systems; netdump, diskdump, LKCD or mcore dumpfiles | package
i | crash-kmp-default | Crash utility for live systems; netdump, diskdump, LKCD or mcore dumpfiles | package
i+ | kdump | Kernel crash dump scripts and utilities | package
i+ | kdumpid | Utility to extract information from vmcores | package
i+ | kexec-tools | Tools for loading replacement kernels into memory | package
i | libkdumpfile10 | Kernel dump file access library | package
i | libKF5Crash5 | An application crash handler | package
i+ | makedumpfile | Partial kernel dump | package
i+ | yast2-kdump | Configuration of kdump | package
kexec-load.service - load default kernel into the current kernel
Loaded: loaded (/usr/lib/systemd/system/kexec-load.service; enabled; preset: disabled)
Active: inactive (dead)
Docs: man:kexec(8)
