Unable to shutdown with AMD Radeon 7800XT

Hey folks,

TL;DR: Can’t fully turn off my computer, how can I find out what causes this, I suspect the GPU?

I am pretty new to Manjaro, coming from Linux Mint hoping a newer Kernel would fix my issue but unfortunately it didn’t so far.
I’ve been silently reading on this issue for about 10 months now. I’ll try to explain what I think is going on.
Last year I got myself a new computer(I’ll list the components below) and I’m pretty happy about how everything works except for the fact that I can’t shut that thing down. Sleep mode works well and it’s what I use 99% of the time now. I can reboot without any issues and every once in a while I try to shutdown and it looks good in the beginning, like a shutdown is supposed to look, and the screens turn dark but then the computer remains running. In this state I am totally locked out, I can only force turn off the system by pressing and holding the power button.
This is not a Manjaro issue but I decided to stay with Manjaro so this is why I am asking for advice here.
I had the same issue with a pretty old default Kernel on Linux Mint and the issue slightly changed in the meantime.
I did a lot of research on this and I am very sure that my issue is highly related to this:
https://bbs.archlinux.org/viewtopic.php?id=289227
I too had rebootloops at some point but now it’s just a system that becomes unresponsive during shutdown. The guy in that thread apparently solved that problem with a specific Kernel version so that’s why I think this issue is about the Kernel.

Unfortunately I can’t even prove or determine that this is my issue because I lack troubleshooting skills on Manjaro, especially when I have to dive that deep.

While I would love an instant solution I mostly came here for troubleshooting advice.
I know there are logfiles about the activities during the shutdown process as I have seen them on other forums. Can anyone help me how to find these logfiles?

I am so sure it’s something about the GPU but of course that’s just an assumption based on issues other people had, so I really would like more information.

I’m running a pretty much up to date Manjaro with the Kernel Version 6.9.5-1, KDE 6

These are the components in use:

CPU:
AMD Ryzen 7 7700X

GPU:
16GB PowerColor Radeon RX 7800 XT Red Devil

Mainboard:
ASRock B650M PRO RS AM5

Cooler:
Endorfy Fortis 5 Cooler

SSD
2TB Lexar NM790

RAM
32GB (2x 16GB) G.Skill Flare EXPO X5 DDR5-6000 DIMM

PSU:
750 Watt be quiet! Pure Power 12 M Modular 80+ Gold

Oh and I hate to admit this but as part of troubleshooting I tried Windows 10 as well and everything worked fine there, so I am confident that it’s not a faulty piece of hardware.

Kind regards
kismet

What if you install kernels 6.6 and 6.1 and try them? Select the kernel to boot on (not the fallback version) in grub at boot.

Everything is better than a hard shutdown right? I would at least try to wake up your PC with:

Maybe reactivate your system with crlt+alt+f3 for TTY (hope it works) and crlt+alt+f2 for back to desktop or try to shutdown from TTY with shutdown commands like:
shutdown -P now

There are other shutdown commands available too.

Btw. since how long you experience this shutdown problems?

Removing Plymouth with deleting this both commands:
quiet splash from Grub, this should atleast show you the steps what your system is doing before its shutdown and probably, where it stucks.

And after saving the Grub file, don’t forget this command:
sudo grub-mkconfig -o /boot/grub/grub.cfg

Here is the Tutorial, but that info that i gave you should already be enough:

1 Like

You can view logfiles with journalctl. Use the --boot=-1 option to view the log from the previous boot. So on the next start after a failed shutdown

journalctl --boot=-1 --no-pager

will output the whole log from the previous boot. Look for anything suspicious towards the end where it’s shutting down. Or upload it to one of these pastebin services and post the link here. E.g.

journalctl --boot=-1 --no-pager | curl -F 'file=@-' https://0x0.st

You could also try

journalctl --boot=-1 --no-pager | grep watchdog

to see any messages about watchdog timers as that seems to have been the issue for some people.

I came from earlier Kernel versions and tried a bunch but I didn’t take specific notes. I will try those again and share the results here.

Unfortunately that didn’t work.

I’ve been dealing with this the entire time, waiting for kernel updates for compatibility since this is a relatively new GPU. The only moment when the shutdown worked was when I put a windows drive in there instead of manjaro or linux mint.

Oh nice I didn’t know how to turn that off! The results look kind of good so far though.

That’s awesome. I did a reboot and tried to shut down quickly afterwards, so all the timestamps are unfortunately pretty close I think… I waited for about 3 minutes after the screens turned black and that did a hard shutdown.
These are the results:
https://0x0.st/XaDt.txt

Watchdog seems to be okay and not the issue here unfortunately.

Doing that will likely leave you in an even worse state.

However, there is a comparatively safer way:

Cheers.

2 Likes

hey @kismet - do you have any user scripts that run on shutdown? check system settings > autostart > logout scripts

if so, delete them and test shutdown and if it works, just re-add them

i saw something in the changelog for systemd that affects this stuff and perhaps that change has propagated to manjaro

possibly related: System won’t shutdown (2024-07-01 update) - Support / System Updates - Manjaro Linux Forum

You should evade hard shutdown on all cost, as @soundofthunder mentioned, to activate reisub is a very good idea.

But as i wrote already, you can also use different (and there are a few) shutdown Terminal commands.

See it as work around till the Startmenue shutdown is working again.

Can you specify entire time?

Nice one, I did try REISUO but the result is the same, the machine remains running. I mean screens are black and it’s not responding to anything but according to a Power measurement plug it’s still pulling about 200W -.-

Sorry I wasn’t more specific. I had to check the receipts to be sure but I assembled the machine in November 2023. Started with Linux Mint and a 5.x Kernel and immediately had that issue. Tried newer Kernels and eventually switched to Manjaro for more up to date stuff.

There is no obvious error there but you do have the sp5100_tco watchdog that is noted as a possible problem in the linked thread from your original post. So first thing to do is try blacklisting it. This won’t do any harm even if it isn’t the cause, I blacklisted it on my system years ago.

sudo sh -c 'echo "blacklist sp5100_tco" > /etc/modprobe.d/disable_sp5100.conf'

and reboot.

If that doesn’t help the next thing I’d look at is Plymouth which seems to cause numerous issues. Try disabling it temporarily by booting to the grub menu, press e, and remove the “splash” option. If that fixes the problem you can remove Plymouth permanently - How to remove plymouth? - #14 by jamjar

If neither of those steps helps then report back.

Done. I blacklistet sp5100_tco and also disabled Plymouth according to the link you provided, I don’t need it anyway and I enjoy having a bit more transparency on what’s happening.
Unfortunately the issue persists. I created a new logfile:
https://0x0.st/XanR.txt

And also because I started to question my memories of everything working with windows I created a WinPE live stick to test and the shutdown works perfectly fine there.

Edit: to all of you, already thank you for your support so far, I already learned a lot today.

To rule out a configuration problem specific to your user account;

Does the shutdown hang happen if you boot from a live Manjaro KDE iso image? If not - create a new user, reboot, login to the new user and test shutdown from there.

Also, I know you said earlier you’d used other kernel versions but do try the 6.6 LTS kernel. I don’t think it’s worth trying anything older than 6.6 with a 7800XT.

Sooo I tried the Manjaro KDE Live ISO and it has a 6.9 Kernel and I get the same result, no working shutdown.

However switching to the 6.6 LTS Kernel did something. With 6.6 I get a reboot instead of a shutdown hang, so it doesn’t crash but reboots right away.
REISUO however gives me the shutdown hang even with 6.6.

That’s interesting because that does exactly match previous reported issues which there have been various attempts at fixing since 6.7.

From what I can make out it’s something to do with power limit settings but I’m not certain. Some people are reporting that reverting this one-line change that got added in 6.8.6 fixes their problem.

But there appear to have been some further changes in 6.10 so the first thing I’d try is installing the 6.10rc4 kernel and see if that helps at all.

If not then I’m afraid further troubleshooting is going to require building and installing a patched kernel…

I tried 6.10rc4 and it seems like it messed with my beloved sleep mode. I can’t properly put the system to sleep anymore, it results in a hang too. Shutdown results in a shutdown hang too, so that gives me the worst experience so far.

That’s some very new ground for me. Is it like taking the source, changing that one line, compiling it? I’ll dig into that.

Can I file an issue or so for this so it will at some point be a regular part of the kernel?

Unfortunately the issue persists on newest (09.10.2024) Manjaro KDE on all kernel versions I’ve tried. I’m bumping this post in hope of a fix.

R5 7600
AsRock B650m Pro RS
AMD RX 7900XTX Reference

Nope, your issue with the newest release update is another story:

Take always a look at the second post (wiki) from the latest stable update.
And you see a overview from the most frequently problem’s.

1 Like

Thank you for your reply.

I do believe that I have the same issue. I have tried to shutdown with “Start with an empty session” but it yields to the same result.

systemctl poweroff

results in the same behaviour as going to the shutdown option in Plasma, that is, DE closes itself, screens go blank, then output from the GPU stops. The PC stays in a hung up state indefinitely without an output.

The GPU in Manjaro Settings Manager is listed under “Unknown device name (ATI Technologies Inc)”

Here is an excerpt from journal logs (journalctl --boot=-1 --no-pager and journalctl --boot=-1 --no-pager | grep watchdog)

After the system hang I’ve innitiaded a reboot with a mobo reset switch.

paź 10 11:45:50 Luna-PC systemd[1]: user.slice: Consumed 5min 12.920s CPU time, 12.7G memory peak.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped target Socket Units.
paź 10 11:45:50 Luna-PC systemd[1]: cups.socket: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Closed CUPS Scheduler.
paź 10 11:45:50 Luna-PC systemd[1]: dirmngr@etc-pacman.d-gnupg.socket: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Closed GnuPG network certificate management daemon for /etc/pacman.d/gnupg.
paź 10 11:45:50 Luna-PC systemd[1]: Removed slice Slice /system/dirmngr.
paź 10 11:45:50 Luna-PC systemd[1]: gpg-agent-browser@etc-pacman.d-gnupg.socket: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Closed GnuPG cryptographic agent and passphrase cache (access for web browsers) for /etc/pacman.d/gnupg.
paź 10 11:45:50 Luna-PC systemd[1]: Removed slice Slice /system/gpg-agent-browser.
paź 10 11:45:50 Luna-PC systemd[1]: gpg-agent-extra@etc-pacman.d-gnupg.socket: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Closed GnuPG cryptographic agent and passphrase cache (restricted) for /etc/pacman.d/gnupg.
paź 10 11:45:50 Luna-PC systemd[1]: Removed slice Slice /system/gpg-agent-extra.
paź 10 11:45:50 Luna-PC systemd[1]: gpg-agent-ssh@etc-pacman.d-gnupg.socket: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Closed GnuPG cryptographic agent (ssh-agent emulation) for /etc/pacman.d/gnupg.
paź 10 11:45:50 Luna-PC systemd[1]: Removed slice Slice /system/gpg-agent-ssh.
paź 10 11:45:50 Luna-PC systemd[1]: gpg-agent@etc-pacman.d-gnupg.socket: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Closed GnuPG cryptographic agent and passphrase cache for /etc/pacman.d/gnupg.
paź 10 11:45:50 Luna-PC systemd[1]: Removed slice Slice /system/gpg-agent.
paź 10 11:45:50 Luna-PC systemd[1]: keyboxd@etc-pacman.d-gnupg.socket: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Closed GnuPG public key management service for /etc/pacman.d/gnupg.
paź 10 11:45:50 Luna-PC systemd[1]: Removed slice Slice /system/keyboxd.
paź 10 11:45:50 Luna-PC systemd[1]: sshd-unix-local.socket: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Closed OpenSSH Server Socket (systemd-ssh-generator, AF_UNIX Local).
paź 10 11:45:50 Luna-PC systemd[1]: systemd-hostnamed.socket: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Closed Hostname Service Socket.
paź 10 11:45:50 Luna-PC dbus-broker[710]: Dispatched 10235 messages @ 2(±3)μs / message.
paź 10 11:45:50 Luna-PC systemd[1]: Stopping D-Bus System Message Bus...
paź 10 11:45:50 Luna-PC systemd[1]: dbus-broker.service: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped D-Bus System Message Bus.
paź 10 11:45:50 Luna-PC systemd[1]: dbus.socket: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Closed D-Bus System Message Bus Socket.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped target System Initialization.
paź 10 11:45:50 Luna-PC systemd[1]: proc-sys-fs-binfmt_misc.automount: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Unset automount Arbitrary Executable File Formats File System Automount Point.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped target Local Encrypted Volumes.
paź 10 11:45:50 Luna-PC systemd[1]: systemd-ask-password-wall.path: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped Forward Password Requests to Wall Directory Watch.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped target Local Integrity Protected Volumes.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped target Local Verity Protected Volumes.
paź 10 11:45:50 Luna-PC systemd[1]: systemd-sysctl.service: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped Apply Kernel Variables.
paź 10 11:45:50 Luna-PC systemd[1]: systemd-coredump.socket: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Closed Process Core Dump Socket.
paź 10 11:45:50 Luna-PC systemd[1]: systemd-modules-load.service: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped Load Kernel Modules.
paź 10 11:45:50 Luna-PC systemd[1]: Stopping Network Time Synchronization...
paź 10 11:45:50 Luna-PC systemd[1]: systemd-update-done.service: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped Update is Completed.
paź 10 11:45:50 Luna-PC systemd[1]: ldconfig.service: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped Rebuild Dynamic Linker Cache.
paź 10 11:45:50 Luna-PC systemd[1]: systemd-journal-catalog-update.service: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped Rebuild Journal Catalog.
paź 10 11:45:50 Luna-PC systemd[1]: Stopping Record System Boot/Shutdown in UTMP...
paź 10 11:45:50 Luna-PC systemd[1]: proc-sys-fs-binfmt_misc.mount: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: run-credentials-systemd\x2dsysctl.service.mount: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: systemd-update-utmp.service: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped Record System Boot/Shutdown in UTMP.
paź 10 11:45:50 Luna-PC systemd[1]: systemd-timesyncd.service: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped Network Time Synchronization.
paź 10 11:45:50 Luna-PC systemd[1]: systemd-tmpfiles-setup.service: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped Create System Files and Directories.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped target Local File Systems.
paź 10 11:45:50 Luna-PC systemd[1]: Unmounting /boot/efi...
paź 10 11:45:50 Luna-PC systemd[1]: Unmounting /tmp...
paź 10 11:45:50 Luna-PC systemd[1]: run-credentials-systemd\x2dtmpfiles\x2dsetup.service.mount: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: boot-efi.mount: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Unmounted /boot/efi.
paź 10 11:45:50 Luna-PC systemd[1]: tmp.mount: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Unmounted /tmp.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped target Swaps.
paź 10 11:45:50 Luna-PC systemd[1]: Reached target Unmount All Filesystems.
paź 10 11:45:50 Luna-PC systemd[1]: systemd-fsck@dev-disk-by\x2duuid-9EB7\x2dEDB4.service: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped File System Check on /dev/disk/by-uuid/9EB7-EDB4.
paź 10 11:45:50 Luna-PC systemd[1]: Removed slice Slice /system/systemd-fsck.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped target Preparation for Local File Systems.
paź 10 11:45:50 Luna-PC systemd[1]: Stopping Monitoring of LVM2 mirrors, snapshots etc. using dmeventd or progress polling...
paź 10 11:45:50 Luna-PC systemd[1]: systemd-tmpfiles-setup-dev.service: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped Create Static Device Nodes in /dev.
paź 10 11:45:50 Luna-PC systemd[1]: systemd-sysusers.service: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped Create System Users.
paź 10 11:45:50 Luna-PC systemd[1]: systemd-remount-fs.service: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped Remount Root and Kernel File Systems.
paź 10 11:45:50 Luna-PC systemd[1]: systemd-tmpfiles-setup-dev-early.service: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped Create Static Device Nodes in /dev gracefully.
paź 10 11:45:50 Luna-PC systemd[1]: run-credentials-systemd\x2dtmpfiles\x2dsetup\x2ddev.service.mount: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: run-credentials-systemd\x2dsysusers.service.mount: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: run-credentials-systemd\x2dtmpfiles\x2dsetup\x2ddev\x2dearly.service.mount: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: lvm2-monitor.service: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Stopped Monitoring of LVM2 mirrors, snapshots etc. using dmeventd or progress polling.
paź 10 11:45:50 Luna-PC systemd[1]: Reached target System Shutdown.
paź 10 11:45:50 Luna-PC systemd[1]: Reached target Late Shutdown Services.
paź 10 11:45:50 Luna-PC systemd[1]: systemd-reboot.service: Deactivated successfully.
paź 10 11:45:50 Luna-PC systemd[1]: Finished System Reboot.
paź 10 11:45:50 Luna-PC systemd[1]: Reached target System Reboot.
paź 10 11:45:50 Luna-PC systemd[1]: Shutting down.
paź 10 11:45:50 Luna-PC systemd[1]: Using hardware watchdog 'SP5100 TCO timer', version 0, device /dev/watchdog0
paź 10 11:45:50 Luna-PC systemd[1]: Watchdog running with a hardware timeout of 10min.
paź 10 11:45:50 Luna-PC kernel: watchdog: watchdog0: watchdog did not stop!
paź 10 11:45:50 Luna-PC systemd-shutdown[1]: Using hardware watchdog 'SP5100 TCO timer', version 0, device /dev/watchdog0
paź 10 11:45:51 Luna-PC systemd-shutdown[1]: Watchdog running with a hardware timeout of 10min.
paź 10 11:45:51 Luna-PC systemd-shutdown[1]: Syncing filesystems and block devices.
paź 10 11:45:51 Luna-PC systemd-shutdown[1]: Sending SIGTERM to remaining processes...
paź 10 11:45:51 Luna-PC systemd-journald[472]: Received SIGTERM from PID 1 (systemd-shutdow).
paź 10 11:45:51 Luna-PC systemd-journald[472]: Journal stopped

        ~  journalctl --boot=-1 --no-pager | grep watchdog                                                                                                 INT ✘   
paź 10 11:41:05 Luna-PC kernel: NMI watchdog: Enabled. Permanently consumes one hw-PMU counter.
paź 10 11:41:06 Luna-PC kernel: sp5100-tco sp5100-tco: Using 0xfeb00000 for watchdog MMIO address
paź 10 11:45:50 Luna-PC systemd[1]: Using hardware watchdog 'SP5100 TCO timer', version 0, device /dev/watchdog0
paź 10 11:45:50 Luna-PC kernel: watchdog: watchdog0: watchdog did not stop!
paź 10 11:45:50 Luna-PC systemd-shutdown[1]: Using hardware watchdog 'SP5100 TCO timer', version 0, device /dev/watchdog0
                                                                                

Don’t do that, activate Reisub to evade Data Loss:

Don’t do that, activate Reisub to evade Data Loss:

REISUO while system is working leads to the same hang as I described previously.

While system is in the described hanged state, neither REISUO nor REISUB is working, probably because Linux kernel is not there anymore and machine is left in a powered on state which I assume is due to incorrect power management of the GPU prior to the kernel shut down.

1 Like