[Stable Update] 2025-08-30 - Kernels, ROCm, Firefox, Steam, PHP

A post was split to a new topic: KDE Dolphin now uses black text on a dark background

I encounter amd gpu driver crashes that freeze my system randomly, printing that message in the system log just before the crash:

sept. 04 20:49:58 kernel: [drm:amdgpu_job_submit [amdgpu]] *ERROR* Trying to push to a killed entity

It was with the linux 6.12.44-2 kernel. I tried the 6.6.103-2, but same problem. I now use the 6.17.0rc3-4, and the problem seems to have disappeared.

I found this fix on LKML : Making sure you're not a bot!. Maybe it is related?

As downgrading to 6.6 didn’t fix the problem, whereas I used that kernel before without crash, I suspect the MESA upgrade. Is it possible that recent MESA triggers a bug in the amd drm driver that it didn’t before?

My hardware is Amd Ryzen R5700G with integrated graphic, using KDE / X11.

2 Likes

Just like to give my experience of the ‘flickering screen’ problem with this update.
I have an NVidia laptop from 2016, so an ‘old’ NVidia, and with kernel 6.1 it started flickering. 5.15, 6.1 and 6.12 also flicker. I tried to downgrade mesa, but that just resulted in a blank screen, couldn’t even ssh. This is even with all nvidia components removed (the laptop also has an Intel ‘card’).
The only thing that stopped the flickering was to switch to kernel 6.15 (which I did accidentally, as mhwd-kernel doesn’t indicate which are are LTS).
I haven’t put NVIdia back in yet but this is the only arrangement that works (and no flicker). Mesa is latest.
So I don’t think it’s got anything to do with NVidia drivers, or possibly even Mesa. If you’re stuck with it, try kernel 6.15

6.15 is EOL. If you want nonlts, use 6.17

1 Like

Just tested it, 6.17 also works without flicker, but still not tried NVidia drivers.

not sure what session/de you are using, but I found flicker on my older hdmi connected monitor in kde plasma / wayland … somehwere in the plasma 6 progression they introduced VRR (variable refresh rate)… once I disabled VRR, flickering no longer happened (typically had occurred after opening/closing full screen games)

It’s a laptop, with Xfce4. X-Windows has nothing to do with it, the flickering starts from the modeset at boot, just after grub. It flickers even if X doesn’t start, even if NVidia drivers are not installed. The only thing that made it go away was a 6.15+ kernel.

The X-Window System, or just “X11” or “X”, please. :grinning:

3 Likes

Want to report an issue with 6.16.x and amdgpu.

After upgrading to 6.16.x I noticed my Journalctl logs are being spammed with amdgpu entries on the power limits being out of range:
kernel: amdgpu 0000:28:00.0: amdgpu: New power limit (213) is out of range [163,200]

Went back to 6.12.44 to test. On 6.12.44 amdgpu generates no entries regarding power limit being out of range at all. The set limit 213 is also maintained.

GPU info:

Graphics:
  Device-1: Advanced Micro Devices [AMD/ATI] Navi 22 [Radeon RX 6700/6700
    XT/6750 XT / 6800M/6850M XT] driver: amdgpu v: kernel
  Display: wayland server: X.org v: 1.21.1.18 with: Xwayland v: 24.1.8
    compositor: kwin_wayland driver: X: loaded: amdgpu
    unloaded: modesetting,radeon dri: radeonsi gpu: amdgpu
    resolution: 3440x1440~145Hz
  API: EGL v: 1.5 drivers: kms_swrast,radeonsi,swrast
    platforms: gbm,wayland,x11,surfaceless,device
  API: OpenGL v: 4.6 compat-v: 4.5 vendor: amd mesa v: 25.2.1-arch1.4
    renderer: AMD Radeon RX 6700 XT (radeonsi navi22 LLVM 20.1.8 DRM 3.61
    6.12.44-3-MANJARO)
  API: Vulkan v: 1.4.321 drivers: radv surfaces: N/A
  Info: Tools: api: clinfo, eglinfo, glxinfo, vulkaninfo
    de: kscreen-console,kscreen-doctor gpu: amdgpu_top,corectrl wl: wayland-info
    x11: xdpyinfo, xprop, xrandr

It seems something got backported from the 6.16.x/6.15.x lines to 6.12.x because the random GPU hard freezes started happening with my 7900 XTX on the LTS kernel as well.

A post was split to a new topic: SMPlayer version 10369 messes up the resolutionsof

Since the last update I cannot send mails with kmail any more… I did not change anything else… anybody an Idea?
I use SMTP and Kmail used to ask me for my password, since I did not store it an then the mail was sent. Now the password is not asked any more…

kmail now relies on a password manager. By default, it’ll look for kwallet, but you can also use keepass{,xc}.

1 Like

Thanks, that helped! There was also an issue with the Akonadie-Console… the Agend was not running properliy. Now all fixed! Thanks!!

1 Like

Hey there, did you install the update by now? Did you experience any issues with the flickering/blackscreen? I am also hesitating to update, since I also have an “old” laptop (2015) and NVIDIA graphic card with nouveau driver…
I am on Kernel 6.1. als well…
wondering, if I should take precautions…maybe I should install a newer kernel before/with the update? would you recommend it?

After upgrade, but before reboot, install 6.15+ kernel as this was the only thing that stopped it for me. Remember, the flickering affects even a tty.
Having another computer handy to ssh in might be a good idea.

2 Likes

After the update my nvidia-containers refused to start.

There happened to be some changes the way the nvidia runtime gets configured.

I found the solution in the archlinux forum:
https://bbs.archlinux.org/viewtopic.php?pid=2258432#p2258432

1 Like

Thanks for the reply and advice! Hmmm…I just checked and found that all kernels above 6.15 are without LTS…that makes me hesitate…
So I don’t really know what to do…probably just not update for now :confused:

I’ve not to taken this or the previous update… mainly based on the “black screen” at boot some folks were reporting, and reports of freezing on the desktop… both having reports coming from nvidia and amdgpu users.

It’s been noted that between the kernel, kernel back-ports, mesa, firmware, and driver (kernel for amd) that there is “something” going on… and AFAIK this sort of implies a “needle in a haystack” issue as we all have varying ages/types of hardware that it would be near impossible to test all scenario’s… nor is there likely to be a “one solution fits all”.

I think what this scenario implies is that “we”/Manjaro don’t have a wide enough sampling of hardware in the unstable/testing branches… I watch those announcement threads as well and I do not recall seeing the volume of reported issues like I’ve been seeing in stable lately. That said, each Manjaro user must pick where their knowledge/comfort level lies… so it’s kind of a “between a rock and a hard place” issue.

However, concerning the kernel back-ports… I thought the core focus was security patches… learning that drivers are also being back-ported kind of creates a condition where no “older kernel” is best for “older hardware”… I feel we are losing snapshots in time… that the well is being poisoned.

1 Like

A reminder that kernel 6.15 reached EOL a few weeks ago, which means updates are no longer provided from the official Manjaro repositories.

Kernels 6.12 (LTS) and 6.6 (LTS) and even 6.1 (LTS) are generally suggested, dependent on your hardware.