Post update I have had 2 of these crashes happen on my machine. Testing the solution now, for anyone else getting random crashes worth having a look at if you’re on AMD.
amdgpu NULL-deref crash on kernel 7.1.3 (RX 9070 XT, ROCm)
Bug
NULL pointer dereference in amdgpu_hmm_invalidate_gfx, triggered during kernel
memory compaction (kcompactd0) migrating a page registered by amdgpu via HMM.
Repeated Oopses cascade into a fatal kernel page fault and hard hang.
Affected kernels: 7.1 and 6.18
Introduced by: drm/amdgpu: fix waiting for all submissions for userptrs
Fix: drm/amdgpu: fix check in amdgpu_hmm_invalidate_gfx
commit 631849ff5d603841e74f19f4a5e30fe1f7d7cf30 (not yet backported to stable 7.1 as of 2026-07-04)
Reduce the compaction/migration path that triggers the bug (runtime + persistent, reversible).
Not a fix - the driver bug remains until commit 631849ff lands.
# /etc/sysctl.d/99-amdgpu-compaction-workaround.conf
vm.compaction_proactiveness = 0
# /etc/tmpfiles.d/thp-defrag.conf
w /sys/kernel/mm/transparent_hugepage/defrag - - - - never
Apply now without reboot:
sudo sysctl -w vm.compaction_proactiveness=0
echo never | sudo tee /sys/kernel/mm/transparent_hugepage/defrag
Status: applied 2026-07-11, under observation.
Edit: Workaround was semi-successful. I got to 34 hours of uptime before a crash. I have now jumped up to the 7.2 experimental kernel. Ran some stress testing, and it all passed, so I will probably stay here until the fix lands in 7.1 or 7.2 goes stable.
I just had a minor issue with the slick greeter: I found /etc/lightdm/slick-greeter.conf.pacsave; in this config file slick-greeter.conf I had set the wallpaper, the font, the mouse cursor size and other things.
Seems that now, the slick greeter use the configuration file /usr/share/lightdm/lightdm.conf.d/90-slick-greeter.conf.
But instead of edit such file, I restored the usual config file, the /etc/lightdm/slick-greeter.conf.pacsave:
/etc/lightdm/slick-greeter.conf: now the greeter looks again with my customizations.
If I should follow another method, please let me know.
KDevelop cannot find libprocesscore.so.11 and crashes when clicking the Configure KDevelop... menu item.
My system only has libprocesscore.so.10, it seems some packages’ version does not match KDevelop’s minimum requirement
kf.coreaddons: "Could not load plugin from /usr/lib/qt6/plugins/kdevplatform/65/kdevgdb.so: Cannot load library /usr/lib/qt6/plugins/kdevplatform/65/kdevgdb.so: libprocesscore.so.11: cannot open shared object file: No such file or directory"
kdevplatform.shell: Can't load plugin "kdevgdb" because a factory to load the plugin could not be obtained: "Could not load plugin from /usr/lib/qt6/plugins/kdevplatform/65/kdevgdb.so: Cannot load library /usr/lib/qt6/plugins/kdevplatform/65/kdevgdb.so: libprocesscore.so.11: cannot open shared object file: No such file or directory"
kf.coreaddons: "Could not load plugin from /usr/lib/qt6/plugins/kdevplatform/65/kdevheaptrack.so: Cannot load library /usr/lib/qt6/plugins/kdevplatform/65/kdevheaptrack.so: libprocesscore.so.11: cannot open shared object file: No such file or directory"
kdevplatform.shell: Can't load plugin "kdevheaptrack" because a factory to load the plugin could not be obtained: "Could not load plugin from /usr/lib/qt6/plugins/kdevplatform/65/kdevheaptrack.so: Cannot load library /usr/lib/qt6/plugins/kdevplatform/65/kdevheaptrack.so: libprocesscore.so.11: cannot open shared object file: No such file or directory"
kf.coreaddons: "Could not load plugin from /usr/lib/qt6/plugins/kdevplatform/65/kdevlldb.so: Cannot load library /usr/lib/qt6/plugins/kdevplatform/65/kdevlldb.so: libprocesscore.so.11: cannot open shared object file: No such file or directory"
kdevplatform.shell: Can't load plugin "kdevlldb" because a factory to load the plugin could not be obtained: "Could not load plugin from /usr/lib/qt6/plugins/kdevplatform/65/kdevlldb.so: Cannot load library /usr/lib/qt6/plugins/kdevplatform/65/kdevlldb.so: libprocesscore.so.11: cannot open shared object file: No such file or directory"
Everything seemed to have went fine, but my keyboard layout was set to US. Set it to my language in KDE settings, but unsure if i would need to set it to my language in other places as well?
~ localectl status ✔
System Locale: LANG=fi_FI.UTF-8
VC Keymap: fi
X11 Layout: (unset)
~ setxkbmap -query ✔
WARNING: Running setxkbmap against an Xwayland server
rules: evdev
model: pc105
layout: us
Isn’t light dm the login manager from Gnome Desktop (or was it XFCE)?
Maybe a mod should redirect that info to made it clear, not to confuse the unexperience people who are using other DE’s and are unaffected like KDE for example.
i updated using pacman, there where some error messages mentioning v4l… After update, i couldn’t start anymore, grub couldn’t find a kernel. I changed-root from USB stick and reinstalled kernel 6.18. That did the trick.
BTW: is there a log of the last update so - next time (hopefully never) - i can give a more specific description.
Unfortunately, I can seem to get into the TTY, I do have a back up Manjaro on a stick but don’t want to go through the long path of repairing or re-installing
Now I was able to get into the TTY, I’m trying some alternatives to try and get things to load
I was on 6.18.35 and trying to manual login to 6.12.95. Will not accept any password I use
I also have this problem with us keyboard set after update and reboot.
Tried this but still us keyboard.
localectl status
System Locale: LANG=sv_SE.UTF-8
VC Keymap: sv-latin1
X11 Layout: (unset)
setxkbmap -query
WARNING: Running setxkbmap against an Xwayland server
Error loading new keyboard description
localectl set-x11-keymap "se"
setxkbmap se
WARNING: Running setxkbmap against an Xwayland server
localectl status
System Locale: LANG=sv_SE.UTF-8
VC Keymap: sv-latin1
X11 Layout: se
So the issue is still there. How to set the Xwayland environment locale?
Nevermind a simple reboot did it. My internet was sluggish as well, anyway weird issue.
I’m not certain whether i understand. In any case the purpose of trying to log into the TTY is to figure out whether the system booted correctly. If you can login then it did. In this case there is a functional Linux system and the black screen only means something in the graphics stack is broken. I would recommend to open a separate thread, post the output of inxi -zv8,. The last part of /var/log/pacman.log (that is the logs of the last update process) could also help to figure out what is wrong. And journalctl - you have full access to everything. The only thing is there is no GUI
So it definitely still recognizes the linux system, I can select from a number of kernels including 6-18.38-1 and 6-12.95 but those both lead to black screens. 6.6.144 also loaded and goes to black screen.
I can access the TTY Menu still which allows me to get to a Grub command line, would I be able to change the video driver reference from there? Or perhaps with editing the boot options?
Ok, yes, I was searching how i could load one of my Timeshift backups from a live USB.
Thankfully i was able to bounce back to an early Timeshift save .. it is a bit dated .. so now I have to be cautious on how I walk the system back up to the recent starting line
-Refreshing keys in order to get Pacman updates working again
Does it actually solve the problem for you? My keyboard is also Finnish but I don’t have any problems even if the output of the setxkbmap -query command is the same as yours.
Update went smoothly, no hiccups at all.
My system started with XFCE, which I still have, but I later added KDE desktop, which I use daily (click my avatar circle to see my other system parameters).
