You are not logged in.
First time I've ever had a crash during a system upgrade and when I rebooted nothing worked so I started manually reinstalling all of the packages that were a part of the last upgrade. It finally let me login, but now I'm getting intermittent and random system freezes.
The mouse stops responding, no caps lock, if audio is playing it just plays the last 2 seconds in a loop. Can't Ctrl-Alt-F2 to switch to another tty. All I can do is reboot.
journalctl -b-1 shows no warnings or errors even with an hour of the freeze.
Here's what I've tried:
* Disabled all non-essential "/' mountpoints to ensure it's not an external disk issue
* Tried latest kernel and linux-lts (6.18.39-1)
* Killed any non-essential programs while running (steam, qbittorrent, conky) other than remmina (for work) and librewolf
* Upgraded the system everytime I have to reboot
* Reinstalled any amd* related pacman pkgs for video drivers
* Installed the pactime aur package, sorted by time and reinstalled EVERY single pkg in the list, so essentially the entire system
What else can I try? The system was rock solid up until this update, I think it was a power surge that caused the crash, it was during some thunderstorms. Other than re-installing the system, I'm not sure what else I can try or investigate, I doubt the freeze is leading to anything being logged to a file or the journal.
Offline
sudo LC_ALL=C pacman -Qkk | grep -v ', 0 altered files' | grep -v backupThough I suspect the subject confuses cause and effect - can you keep "dmesg -w" running in a permanently visible terminal? Maybe is shows something before the kernel halts.
Online
Oh yeah, forgot to list that I already went through the -Qkk and reinstalled all of those:
warning: amd-ucode: /boot/amd-ucode.img (Permissions mismatch)
warning: cups: /etc/cups/classes.conf (Permissions mismatch)
warning: cups: /etc/cups/printers.conf (Permissions mismatch)
warning: libutempter: /usr/lib/utempter/utempter (GID mismatch)
warning: libutempter: /usr/lib/utempter/utempter (Permissions mismatch)
warning: libvlc: /usr/lib/vlc/plugins/plugins.dat (Modification time mismatch)
warning: libvlc: /usr/lib/vlc/plugins/plugins.dat (Size mismatch)
warning: libvlc: /usr/lib/vlc/plugins/plugins.dat (SHA256 checksum mismatch)
warning: openssl: /etc/ssl/private (Permissions mismatch)
warning: systemd: /var/log/journal (GID mismatch)
amd-ucode: 17 total files, 1 altered file
cups: 947 total files, 2 altered files
libutempter: 20 total files, 1 altered file
libvlc: 120 total files, 1 altered file
openssl: 6748 total files, 1 altered file
systemd: 1918 total files, 1 altered filedemsg is running now, waiting for a freeze.
Offline
Another freeze. Happened literally the same second I clicked in the browser. Nothing in dmesg, nothing in journalctl -b-1. Updated my bios at the reboot.
Offline
Another freeze. Happened literally the same second I clicked in the browser. Nothing in dmesg, nothing in journalctl -b-1. Updated my bios at the reboot.
I really want to see how that would be solved, try this https://wiki.archlinux.org/title/Keyboard_shortcuts, maybe when the system freezes you hard reset it (using button shutdown), I think you can try shutting down sepecific things to check. {even though I have no experience on that}
I just remembered something similar happened to my system for about 7 seconds after opening the browser (don't really remember what else} three days ago, but everything returned to normal afterwards so I ignored it {thought it may be memory related but didn't check}, I'm using dwm and as mpv was playing it started to loop until the end of the crash, I update my system daily btw cause I have a fast connection ![]()
Last edited by aminepro (2026-07-23 20:59:45)
Offline
Specifically https://wiki.archlinux.org/title/Keyboa … el_(SysRq) but if there're no errors in the live dmesg that's not even overly likely.
=> https://wiki.archlinux.org/title/Kdump
Still post your system journal for a general oversight so we can round up the usual suspects, eg. https://wiki.archlinux.org/title/Ryzen#Troubleshooting
Are you overclocking the system in any way shape or form? (XMP etc?)
Online
Nope, no overclocking, no custom kernels. All was perfect until a borked pacman -Syu.
I did see the System Halts section, but I'm not sure what PBO means in the context of: "then apply the PBO+4 curve for more stability"
Also installed kdumpst to see if I can get a dump at next freeze.
Offline
Do you have a ryzen CPU?
Please post your complete system journal for the boot:
sudo journalctl -b | curl -s -H "Accept: application/json, */*" --upload-file - 'https://paste.c-net.org/'(Just to illustrate the system we're facing)
Online
There's the APU and a Radeon RX 9060 XT, can you run the system on the integrated graphics for a while and see whether that has impact on the stability?
ASUS System Product Name/PRIME X870-P WIFI, BIOS 1644 03/10/2026Not sure whether it does (eg. the TUF stuff seems to have castrated UEFI options) but the PBO curve optimizer is a BIOS/UEFI option - you're basically running the CPU at higher voltages what keeps it "warmer" and prevents it from overdriving isolated cores toomuch™
Disabling the c-states might have similar impact, but at the cost of quite higher power draws (and it's not as reliable)
There's also https://archlinux.org/packages/extra/x86_64/corectrl/ (you *want* to use the UEFI settings if they're available)
If it's not the CPU/GPU, for the moment just ftr, there's a Micron/Crucial Technology T500 NVMe
Online
I've never actually used the integrated graphics because I have to run 2 monitors, so that's out. I looked in the bios for power management features but the only things I saw were Wake On Lan and 1 other thing that now I can't remember, but they were both disabled.
The NVME reports no smart errors in the bios SMART util.
Offline
https://www.asus.com/de/motherboards-co … 70-p-wifi/ makes it look like there're some elaborate settings and certainly PBO control?
Even if it was the nvme that would not relate to smart errors (broken memory modules) but the bus, again: only for later.
Focus on the CPU and related to that the GPU.
Online
Just a quick update, reinstalled my arch system from scratch, reinstalled librewolf from source and while that was compiling, installed steam and associated vulkan-radeon drivers, mesa etc. Literally within 5 minutes of getting that done it locked up again.
So I started over and decided to leave my video drivers alone for a while. Re-installed arch again, this time only installed firefox so I could use the machine. Installed remmina so I could rdp to my work laptop. System seemed pretty stable, played music, surfed the web. Then I came back that evening, it was still up.
Decided to install a few minor utils and some files from backup to my desktop. Did a `sudo pacman -S vlc` and selected all plugins. It froze within 30s and I didn't even start vlc. It could be a coincidence, but now I've uninstalled vlc and I'm leaving my system overnight with no extra gpu drivers and no vlc.
We'll see what happens.
Offline
System seemed relatively stable, installed steam today after enabling the 32-bit multilib. Logged the pkgs installed:
Packages (95) alsa-plugins-1:1.2.12-5 appstream-1.1.5-1 hwloc-2.14.0-1 lib32-alsa-lib-1.2.16.1-1 lib32-alsa-plugins-1.2.12-1
lib32-brotli-1.2.0-2 lib32-bzip2-1.0.8-4 lib32-curl-8.21.0-1 lib32-dbus-1.16.2-1 lib32-e2fsprogs-1.47.4-1
lib32-expat-2.8.2-1 lib32-fontconfig-2:2.18.2-1 lib32-freetype2-2.14.3-1
lib32-gcc-libs-16.1.1+r581+gb73ad535acaa-1 lib32-glib2-2.88.2-1 lib32-glibc-2.44+r3+g0b05bc142249-1
lib32-gmp-6.3.0-2 lib32-gnutls-3.8.13-3 lib32-icu-78.3-1 lib32-json-c-0.19-1 lib32-keyutils-1.6.3-4
lib32-krb5-1.22.2-1 lib32-libdisplay-info-0.3.0-1 lib32-libdrm-2.4.134-1 lib32-libelf-0.195-1
lib32-libffi-3.7.1-1 lib32-libgcrypt-1.12.2-1 lib32-libglvnd-1.7.0-1 lib32-libgpg-error-1.61-1
lib32-libidn2-2.3.8-1 lib32-libldap-2.6.13-1 lib32-libnghttp2-1.69.0-1 lib32-libnghttp3-1.18.0-1
lib32-libngtcp2-1.25.0-1 lib32-libnm-1.58.0-1 lib32-libpciaccess-0.19-1 lib32-libpipewire-1:1.6.8-1
lib32-libpng-1.6.58-2 lib32-libpsl-0.21.5-1 lib32-libssh2-1.11.1-4 lib32-libtasn1-4.21.0-1
lib32-libunistring-1.4.2-1 lib32-libva-2.24.1-1 lib32-libx11-1.8.13-1 lib32-libxau-1.0.12-1
lib32-libxcb-1.17.0-1 lib32-libxcrypt-4.5.2-1 lib32-libxcrypt-compat-4.5.2-1 lib32-libxdmcp-1.1.5-1
lib32-libxext-1.3.7-1 lib32-libxfixes-6.0.1-2 lib32-libxinerama-1.1.6-1 lib32-libxml2-2.15.3-1
lib32-libxshmfence-1.3.3-1 lib32-libxss-1.2.5-1 lib32-libxxf86vm-1.1.5-2 lib32-llvm-libs-1:22.1.8-1
lib32-lm_sensors-1:3.6.2-2 lib32-mesa-1:26.1.5-1 lib32-ncurses-6.6-2 lib32-nettle-4.0-2 lib32-nspr-4.39-1
lib32-nss-3.126-1 lib32-openssl-1:3.6.3-1 lib32-p11-kit-0.26.4-1 lib32-pcre2-10.47-1 lib32-pipewire-1:1.6.8-1
lib32-spirv-tools-1:1.4.350.1-1 lib32-sqlite-3.53.4-1 lib32-systemd-261.2-1 lib32-util-linux-2.42.2-1
lib32-vulkan-icd-loader-1.4.350.1-1 lib32-vulkan-mesa-implicit-layers-1:26.1.5-1 lib32-vulkan-radeon-1:26.1.5-1
lib32-wayland-1.25.0-1 lib32-xcb-util-keysyms-0.4.1-2 lib32-xz-5.8.3-1 lib32-zlib-1.3.2-1 lib32-zstd-1.5.7-2
libadwaita-1:1.9.2-1 libblake3-1.8.4-1 libfyaml-0.9.6-2 libxcrypt-compat-4.5.2-1 libxmlb-0.3.29-1
lsb-release-2.0.r55.a25a4fc-1 lsof-4.99.7-1 onetbb-2023.1.0-1 steam-devices-1.0.0.87-1 usbutils-019-1
vulkan-mesa-implicit-layers-1:26.1.5-1 vulkan-radeon-1:26.1.5-1 xdg-user-dirs-0.20-1 xorg-xrandr-1.5.4-1
zenity-4.2.2-1 steam-1.0.0.87-1Got a freeze within the hour while watching a youtube video.
So if I try reinstalling steam and instead of selecting vulkan-radeon I select vulkan-nouveau, would that use the generic drivers instead ruling out amdgpu specific issues?
Offline
There's no nvidia card in the journal you posted, you can try using swrast or swiftshader, https://wiki.archlinux.org/title/Vulkan#Installation
But
1. there're two AMD GPUs
Jul 24 10:07:34 archlinux kernel: amdgpu 0000:75:00.0: [drm] Cannot find any crtc or sizesThere're no output attached to the APU => just disable it? To check whether the issue is (if at all) coming from the RX 9060
2. "The mouse stops responding, no caps lock, if audio is playing it just plays the last 2 seconds in a loop." - this kind of kernel halt rarely occurs for a graphics driver issue
3. "installed steam and associated vulkan-radeon drivers, mesa etc. Literally within 5 minutes of getting that done" just installing these things will not have had any impact unless you re/started a GUI session
Online
Hmm, well I tried the boot param to disable GFXOFF and that didn't do anything.
Also tried enabling kdumpst and simple-kdump. kdumpst didn't seem to do anything and the journal never showed it load. simple-kdump got setup, but when I tested the kernel panic, it literally just hung like it always does, didn't reboot and when I did reboot manually, there was no crash file. Seems like the wiki might need to be updated for that or I just didn't understand some assumptions.
I also followed through this other guy's thread: https://bbs.archlinux.org/viewtopic.php?id=304412&p=3 which looks eerily similar.
There's also this thread: https://bbs.archlinux.org/viewtopic.php … 0#p2261920 which looks similar, but unless they just stuck the binary blob back in linux-firmware-amdgpu I doubt it regressed again.
I think I'm just going to have to bite the bullet and reinstall and start logging every single pkg that gets installed, ideally 1 per day until I get a crash.
On the plus side, I've gotten a lot more comfortable messing with grub entries and kernel parameters.
Oh yeah, the mobo has integrated graphics which I've never used, I'm not even sure if I can disable it, I'll check the bios at next reboot.
Offline
https://wiki.archlinux.org/title/Kdump#Manual_steps - you don't have to compile a kernel.
Online
I know, I'm running linux-lts which it says is already configured for kdump. But following the text in the wiki for both simple-kdump nor kdumpst led to anything actually happening/being written at panic time.
Offline
written at panic time
That is with a synthetically triggered kernel crash??
https://wiki.archlinux.org/title/Kdump# … the_kernel
What are the literal crashkernel and kexec parameters you used?
(There're a bunch of placeholders in the wiki examples)
Online
I followed both Automatic sections. For simple-kdump I set crashkernel=2G, rebuilt the grub cfg and rebooted into that crashkernel. Then I attempted the panic from:
> Finally enable and start simple-kdump-setup.service, then refer to #Testing kdump by crashing the kernel to verify the kdump behavior.
It just froze like it always does, I had to manually reboot with power switch and the wiki says:
> The kexec kernel should reach target Emergency Mode to collect vmcore, with a prompt asking login for the emergency shell. You can ignore that login as the vmcore collection will happen at the background and reboot automatically.
That didn't happen, it just booted normally and there was no file in /var/crash.
Anyway, I've reinstalled the base Arch system again with limited pkgs just so I can work, Librewolf, remmina and pulseaudio. Freeze still happened while playing a youtube video. I then modified the boot parameters to loglevel=8 and currently waiting for a crash.
Offline
This has nothing to do with what packages are installed and probably also not w/ what processes are runnning
https://www.asus.com/de/motherboards-co … 70-p-wifi/ makes it look like there're some elaborate settings and certainly PBO control?
There's also https://archlinux.org/packages/extra/x86_64/corectrl/ (you *want* to use the UEFI settings if they're available)
Ans at the very least try to limit the c-states to 1, https://wiki.archlinux.org/title/Ryzen#Troubleshooting
Online
I found this thread which looks like he narrowed it down to CPB which I've got currently disabled. If that still crashes, I'll disable PBO next. And then move on to c-state modifications.
I was really leaning towards a pkg issue, because I've been running this desktop 100% stable for over a year until a pacman update with 0 changes to hardware or BIOS settings.
Last edited by KaibutsuX (2026-07-29 20:28:39)
Offline
https://wiki.archlinux.org/title/Arch_L … cific_date
microcode or firmware files might have an impact but the ryzen problem is a degrading condition ie it will get worse over time.
Online
Tried amd-ucode, but it appears the kernel already has the most up to date patch for my cpu without it installed.
Offline
https://wiki.archlinux.org/title/Arch_L … cific_date
microcode or firmware files might have an impact but the ryzen problem is a degrading condition ie it will get worse over time.
Where are you getting that information that it degrades over time?
Offline