You are not logged in.
linux 6.13.6.arch1-1
mesa 1:25.0.1-1
works without complaints and seems stable.
cat /sys/kernel/debug/dri/1/amdgpu_gpu_recover doesn't freeze the system, although it adds errors:
amdgpu: psp gfx command UNLOAD_TA(0x2) failed and response status is (0x117)
[drm:amdgpu_ib_ring_tests [amdgpu]] *ERROR* IB test failed on comp_1.1.0 (-110).
Blame the google-translator for all language errors.
Offline
@OJaksch
llvm 20.1.0 has been released. The llvm and llvm-libs packages will not be long in coming.
As soon as they are released in the arch repositories, you will no longer be able to use the mesa-test-git packages.
https://llvm.org/
Oi, thanks for that hint!
Offline
@OJaksch
llvm 20.1.0 has been released. The llvm and llvm-libs packages will not be long in coming.
As soon as they are released in the arch repositories, you will no longer be able to use the mesa-test-git packages.
https://llvm.org/@lpr
thanks. mesa 25.0.1 from /testing and try it out? Or wait until it is pushed to /Extra.
I\m waiting, I'm not bothering with this nonsense anymore, I don't have nerves for it
Offline
Mesa 25 is on the normal repos now, so everything should be normal without any special versions now.
Offline
linux-lts-6.12.19-1-x86_64
mesa-1:25.0.1-2-x86_64 | vulkan-radeon-1:25.0.1-2-x86_64
llvm-libs-19.1.7-1-x86_64
linux-firmware 20250311.b69d4b74-2
amd-ucode 20250311.b69d4b74-2
Since yesterday mesa 25.0.1-2 is also in the testing repos of Debian, so one can hope that mesa 25.0.x will be included in Debian 13 trixie. Debian 12 uses mesa 22.3.6/24.2.8.
This could also get the cow off the ice for the other distributions.
Since my enquiry [#672] was not answered here, I assume, like @pacmancrashedagain, that the problems discussed here have been solved.
My criticism of the handling of this pervasive bug remains and I have drawn conclusions from it, but they did not begin with this. I am a little worried - Linux is the best existing operating system for me.
Without @Lone_Wolf I would have been lost. Thanks also to @pacmancrashedagain for starting this thread.
Offline
I think this topic can be marked as "solved". After ~half a year of using old versions of mesa and llvm-libs the computer seems stable and free of errors. Thanks to all contributors.
Last edited by Berni341 (2025-03-17 12:34:07)
Offline
...Before that happens, I would have liked to know whether mesa has revised the fix again as announced or will revise it after a fix has also been introduced in the kernel (#649).
Offline
Since the upgrade to Linux 6.13.5, I have not seen any issues. About 14 days stable uptime with 24.3.4-1 mesa, and about 7 days stable uptime with 25.0.1-2 mesa and Linux 6.13.7. Seems solved, knock, knock.
"Although the masters make the rules
For the wise men and the fools
I got nothing, Ma, to live up to."
Offline
I'm just checking if everyone has been stable lately, i've had a couple of lockups since Mesa 25.0.3 and i haven't found anything in the logs, although there's been a new version of the qt6-webengine which looks like it has some issues and i'm using qutebrowser so it might be that.
Offline
I had the problem since the start of this topic, and have no problem since the fixes have been merged. Currently on 25.0.3.
Offline
Stable.
linux 6.13.8.arch1-1
linux-headers 6.13.8.arch1-1
linux-firmware 20250311.b69d4b74-3
glibc 2.41+r9+ga900dbaf70f0-1
mesa 1:25.0.2-2
we are not condemned to write ugly code
Offline
Apparently, the problem is back.
I'm not using Arch Linux, but a distro derived from it, Manjaro.
On 04/13/2025 mesa was updated to version 1:25.0.3-1 and today my laptop crashed in the same way as before. On mesa 25.0.2 everything was fine, I even spent whole days without any problems, today after just over 3 hours a crash,
Offline
I can't reproduce a crash with my reproducer with mesa 1:25.0.3-1 and linux 6.14.2.arch1-1.
Offline
I still have also this problem with Manjaro (testing).
The issue only occurs after the monitors have been turned off by power management.
In other words, when the monitors wake up again, there's a risk that these freezes happen while watching videos.
Still, it seems to be pretty random.
Offline
Manjaro is not Arch and not supported here, ask Manjaro devs.
FWIW since this is a bunch of interaction on the kernel level as well that is going to be relevant, make sure you are on an up to date kernel.
Offline
Hello,
Yes, I know that Arch is not Manjaro, but since distributions derived from other distributions use many of the core packages (Linux Mint/Ubuntu/Debian for example), this was one of the few places where I found an in-depth discussion of the case, including finding here the temporary solution of keeping mesa and other packages at version 24.2.7.
The intention here was just to leave another report that the problem may still be present, but not as common.
Commenting now as a Linux user in general, it is worrying that a problem that affects Kernel 6.12, which is LTS, and AMD cards that have the opensource driver, has passed me by silently. I did not find anything openly discussed, alerts on sites like Phoronix, I found it by searching through forums, and again, the most talked about thing I found was here. I don't know if this was because it probably affected more distros with rolling releases or those who use unstable versions of the distributions, but it is still a concern. At least I couldn't find any material on a website or any guidance from any distro about this problem.
Anyway, I would like to thank everyone who has participated in this topic so far, as the information provided here has been very helpful.
Thank you.
Offline
@ZacharyFoxx and @heebo would you mind describing your hardware configuration?
Last edited by Mechanicus (2025-04-15 17:03:18)
Offline
I don't know if this was because it probably affected more distros with rolling releases or those who use unstable versions of the distributions, but it is still a concern. At least I couldn't find any material on a website or any guidance from any distro about this problem.
Manjaro had held on to mesa 24.2.7/24.2.8 for a long time and was therefore only affected by the bug for a short time before the fix was at least included in the kernel. By the way, Manjaro (Philip Müller) packages mesa themselves, so that it does not completely match the arch package. The reason for this was the removal of hardware acceleration for AMD graphics cards.
My system (arch) has no more problems since the kernel got a fix (since linux 6.13.5 and linux-lts 6.12.17). mesa itself had left us out in the rain for months.
However, I would still like to know whether the mesa fix has been revised later, as it was announced.
Does anyone know anything about this? In the event of a change, the stability would have to be checked again...
Offline
However, I would still like to know whether the mesa fix has been revised later, as it was announced.
Does anyone know anything about this? In the event of a change, the stability would have to be checked again...
MR 33248 with the fix only changed 1 file , src/gallium/drivers/radeonsi/si_pipe.c
Keep an eye on the git history of commits changing si_pipe.c to verify if there have been related changes.
Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.
clean chroot building not flexible enough ?
Try clean chroot manager by graysky
Offline
@ZacharyFoxx and @heebo would you mind describing your hardware configuration?
Of course.
System:
Host: heebo-desktop Kernel: 6.13.9-1-MANJARO arch: x86_64 bits: 64
Desktop: KDE Plasma v: 6.3.4 Distro: Manjaro Linux
Machine:
Type: Desktop Mobo: ASRock model: X300M-STX serial: <superuser required>
UEFI: American Megatrends LLC. v: P2.10 date: 11/21/2024
CPU:
Info: 6-core model: AMD Ryzen 5 4600G with Radeon Graphics bits: 64
type: MT MCP cache: L2: 3 MiB
Speed (MHz): avg: 2381 min/max: 400/4308 cores: 1: 2381 2: 2381 3: 2381
4: 2381 5: 2381 6: 2381 7: 2381 8: 2381 9: 2381 10: 2381 11: 2381 12: 2381
Graphics:
Device-1: Advanced Micro Devices [AMD/ATI] Renoir [Radeon Vega Series /
Radeon Mobile Series] driver: amdgpu v: kernel
Display: wayland server: X.org v: 1.21.1.16 with: Xwayland v: 24.1.6
compositor: kwin_wayland driver: X: loaded: amdgpu unloaded: modesetting
dri: radeonsi gpu: amdgpu resolution: 1: 1920x1080~60Hz 2: 1920x1080~60Hz
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: 25.0.3-arch1.1
renderer: AMD Radeon Graphics (radeonsi renoir ACO DRM 3.61
6.13.9-1-MANJARO)
API: Vulkan v: 1.4.309 drivers: radv surfaces: xcb,xlib,wayland
Info: Tools: api: clinfo, eglinfo, glxinfo, vulkaninfo
de: kscreen-console,kscreen-doctor gpu: amdgpu_top,radeontop
wl: wayland-info x11: xdpyinfo,xprop
Audio:
Device-1: Advanced Micro Devices [AMD/ATI] Renoir Radeon High Definition
Audio driver: snd_hda_intel
Device-2: Advanced Micro Devices [AMD] Family 17h/19h/1ah HD Audio
driver: snd_hda_intel
API: ALSA v: k6.13.9-1-MANJARO status: kernel-api
Server-1: PipeWire v: 1.4.1 status: active
Network:
Device-1: Realtek RTL8111/8168/8211/8411 PCI Express Gigabit Ethernet
driver: r8169
IF: enp2s0 state: up speed: 1000 Mbps duplex: full mac: 9c:6b:00:3e:4c:8b
Device-2: Intel Dual Band Wireless-AC 3168NGW [Stone Peak] driver: iwlwifi
IF: wlp3s0 state: down mac: fa:e2:a0:8d:6d:9b
Bluetooth:
Device-1: Intel Wireless-AC 3168 Bluetooth driver: btusb type: USB
Report: btmgmt ID: hci0 state: up address: F4:26:79:FB:71:8C bt-v: 4.2
Drives:
Local Storage: total: 704.25 GiB used: 184.52 GiB (26.2%)
ID-1: /dev/nvme0n1 vendor: KIOXIA model: EXCERIA G2 SSD size: 465.76 GiB
ID-2: /dev/sda vendor: Crucial model: CT128MX100SSD1 size: 119.24 GiB
ID-3: /dev/sdb vendor: Crucial model: CT128MX100SSD1 size: 119.24 GiB
Partition:
ID-1: / size: 448.43 GiB used: 78.64 GiB (17.5%) fs: ext4
dev: /dev/nvme0n1p2
ID-2: /boot/efi size: 299.4 MiB used: 31.3 MiB (10.4%) fs: vfat
dev: /dev/nvme0n1p1
Swap:
ID-1: swap-1 type: partition size: 8.8 GiB used: 0 KiB (0.0%)
dev: /dev/nvme0n1p3
Sensors:
System Temperatures: cpu: 43.1 C mobo: N/A gpu: amdgpu temp: 37.0 C
Fan Speeds (rpm): N/A
Info:
Memory: total: 16 GiB note: est. available: 13.51 GiB used: 6.1 GiB (45.2%)
Processes: 360 Uptime: 2h 20m Shell: Zsh inxi: 3.3.38
Offline
Keep an eye on the git history of commits changing si_pipe.c to verify if there have been related changes.
thanks a lot!
Offline
@ZacharyFoxx and @heebo would you mind describing your hardware configuration?
Here are:
System:
Host: nitro Kernel: 6.12.21-4-MANJARO arch: x86_64 bits: 64
Desktop: Xfce v: 4.20.1 Distro: Manjaro Linux
Graphics:
Device-1: NVIDIA GA107M [GeForce RTX 3050 Mobile] driver: nvidia
v: 570.133.07
Device-2: Advanced Micro Devices [AMD/ATI] Rembrandt [Radeon 680M]
driver: amdgpu v: kernel
Display: x11 server: X.org v: 1.21.1.16 driver: X: loaded: amdgpu,nvidia
unloaded: modesetting,nouveau dri: radeonsi
gpu: amdgpu,nvidia,nvidia-nvswitch resolution: <missing: xdpyinfo/xrandr>
resolution: 1: 1920x1080 2: 1920x1080
API: OpenGL v: 4.6.0 compat-v: 4.5 vendor: amd mesa v: 25.0.3-arch1.1
renderer: AMD Radeon 660M (radeonsi rembrandt LLVM 19.1.7 DRM 3.61
6.12.21-4-MANJARO)
Info: Tools: api: eglinfo, glxinfo, vulkaninfo de: xfce4-display-settings
gpu: nvidia-settings,nvidia-smi x11: xprop
Offline
The issue is a combination between kernel and mesa, you ideally need both a more recent kernel and a recent mesa. Test a 6.14 kernel. -- or even a newer LTS as well, you're two patch versions out..
Last edited by V1del (2025-04-17 16:51:29)
Offline
Thanks @ZacharyFoxx and @heebo for sharing! Your configurations were not affected neither by the change in the kernel nor by the change in mesa 25. If you can compile the custom kernel, I can provide the fix to try. Or I can provide the kernel build, but since I'm using ArchLinux, I'm not sure it will work flawless on Manjaro.
Offline
The issue is a combination between kernel and mesa, you ideally need both a more recent kernel and a recent mesa. Test a 6.14 kernel. -- or even a newer LTS as well, you're two patch versions out..
Yes, I just find out that there is 6.14 kernel available and installed it. There is also 6.15rc, but left that out for now.
I have no idea how to compile kernel myself, so better leave that for now.
Offline