System crashes/hangs coming out of sleep

Ask Fedora
,
1

A while ago (a couple of months) I had an issue where installing an F40 update appeared to hang my main machine - AMD 7950X, MSI PRO mobo, DDR5-6000 - during the process and I ended up with an unbootable machine.

Since then I have this issue which seems 100% reproducible; it could be a chicken or an egg, I’m not sure which. As usual I searched for a similar problem first, but didn’t find anything recent - there looked to be a similar issue back in 2022.

The machine sleeps, either via timeout, or manually. If I try to bring it back out of sleep it starts to paint the login screen and then hangs. I’m pretty sure it crashes as there’s no response - no keyboard, no mouse. If I do the same on dual-booted Windows 11 everything is AOK so I don’t believe it’s specifically a machine problem, but obviously it could be since I have no diagnostics.

I see this via journalctl just before the “-- Boot” line after I hard reset it but I don’t have any keener insight to what might be going wrong:

Nov 01 16:26:04 fedora systemd[1]: Starting systemd-suspend.service - System Suspend...
Nov 01 16:26:04 fedora wpa_supplicant[1647]: wlp15s0: CTRL-EVENT-DSCP-POLICY clear_all
Nov 01 16:26:04 fedora systemd-sleep[9406]: Performing sleep operation 'suspend'...
Nov 01 16:26:04 fedora kernel: PM: suspend entry (deep)
Nov 01 16:26:04 fedora kernel: Filesystems sync: 0.011 seconds

There are 2 4K displays attached to an MSI RX 6800 graphics card.

The length of time the machine is asleep appears immaterial: a few seconds or several days (I have other machines…)

Anyone got any insight into info I might gather to try and track down what’s happening?

The machine also has Debian installed but when I had to do a full reinstall of Fedora after the last hang during upgrade GRUB got nuked and I can’t currently access Debian until I reinstall it (I guess, or try my hand at grub.conf updating manually; I think reinstalling will be easier - it’s my main dev machine)

Thanks for any thoughts, info or insights.

I always forget something… I’m going to try upgrading to F41, but I’m not anticipating that will fix it. I’m not sure what will.

1 Like
2

It’s probably a unique hardware/software config. I have another machine here (the one I’m typing on now) that’s a 5950X system, MSI mobo with DDR4/128GB, MSI RX 6600 GPU, multi-booted with Windows 11 and several Linux distros and it’s behaving fine. (The 5950X has been rock solid!)

The AMD 7000/670E series has had a ridiculous amount of BIOS updates and it’s quite possible that’s the culprit. I have no diagnostic info so I can’t find out one way or the other. I’d just be interested to know that I’m not the only one…

1 Like
3

Same issue, started happening after update to Fedora 41 (from Fedora 40). Did not look at journalctl.

Machine: Lenovo LOQ Laptop
OS: Fedora 41 (Windows 11 dual boot)
CPU: AMD Ryzen 7 7840HS w/ Radeon 780M Graphics
GPU: NVIDIA GeForce RTX 4060 Laptop

4

Well, another data point: triple-booted machine - Fedora 40, Windows 11 Pro, Debian 12 - and only Fedora is hanging/crashing coming out of sleep. The other two OSs are behaving correctly. I conclude therefore that it’s not a hardware/BIOS issue (or if it is the necessary combination is unique to Fedora 40). Nothing interesting/useful in dmesg/journalctl that I can tell. Of course when the machine hangs I have to hard reboot it so neither dmesg or journalctl can provide any help.

Given the amount of problems that appear to have arisen with Fedora 41 that I’ve seen on the forum in the past couple of days I think I’ll hold off upgrading to 41 for the time being.

In the meanwhile I’ll have to stop the machine going to sleep when running F40 (or, you know, flip into Windows or Debian). I think I may try swapping the 6800 for an RTX 3060, but I shouldn’t have to … (plus all the attendant difficulty getting Nvidia cards to work on Linux should I want to venture further than Nouveau).

5

On one of my laptop, almost always wireplumber and bluez crash after suspend & wake. It started to happen a few months ago. But it doesn’t affect usage besides abrt constantly telling me they crashed.
The laptop is not good so I assume it was happening before, just silently, and now they added something to signal the crash. Also looking at wireplumber Bugzilla, there are a mountain of crash reports, so I guess it’s normal…

6

pipewire was causing a lot of problems for me at one point: there were hundreds or thousands of error messages in journalctl, but it all just seemed to get resolved at some point.

1 Like