I have a 3900x running on an X570 Gigabyte Aorus Master motherboard. Since purchase at release time i’ve never been able to get this system to sleep or hibernate under linux regardless of distro or kernel. I’ve followed all of the suspend troubleshooting steps on the arch wiki to no avail and the most googling has done was find me a couple people with similar issues who never resolved them publicly.
The system will suspend/hibernate then immediately wake 2-3 seconds later. This was also an issue under windows when the board was new however an updated LAN driver for the onboard realtek 2.5g nic solved it. Disabling the nic in the BIOS curiously had no effect on this problem.
If anyone has managed to solve this i’d be very grateful for any tips. I’m trapped on windows because of this until I can purchase new hardware due to this bug and it sucks.
I have had this problem with a few solutions that worked.
-
A mouse I had (drivers or something) kept waking it up. I disabled the mouse’s ability to wake up the PC.
-
The feature that turns off the HDD was interfering with the sleep mode. I had to make sure the HDD turns off BEFORE the PC sleeps.
I just read you are on linux. I should have read more carefully.
Well, those were two issues I had. It was in Windows though.
Can a Linux-savy person say if there is something like “powercfg.exe -lastwake” on Linux, this way you could poke at this issue a bit more focussed.
I’ve recently had a similar issue I’ve had to review - in this case Ubuntu, (refusing to sleep after 00:00, very irritating).
Arch resources led to isolating a cause, in my case; to credit the reference:
https://bbs.archlinux.org/viewtopic.php?pid=1823343#p1823343
# turn on trace
echo 1 | sudo tee /sys/power/pm_trace
# intiate a suspend
sudo /usr/lib/systemd/systemd-sleep suspend
# when returned, scan syslog for `PM:’
less /var/log/syslog
In my case it led to fsutil. and to destroy the process with prejudice (it’s before 1am, on a weekday) and suspend works.
In hope this may be of use.
Don’t forget to undo the trace (same command with 0)
1 Like
Thank you everyone for the help. The logs don’t show a whole lot from what I can gather?
Here is the output from journalctl after setting pm_trace to 1 and grepping for PM:
Jul 24 13:06:21 diablo-x570aorusmaster kernel: PM: Enabling pm_trace changes system date and time during resume.
PM: Correct system time has to be restored manually after resume.
Jul 24 13:06:26 diablo-x570aorusmaster kernel: PM: suspend entry (deep)
Jul 24 13:06:32 diablo-x570aorusmaster kernel: PM: Saving platform NVS memory
Jul 24 13:06:32 diablo-x570aorusmaster kernel: PM: Restoring platform NVS memory
Jul 24 13:06:32 diablo-x570aorusmaster kernel: ptp ptp1: PM: parent enp10s0 should not be sleeping
Jul 24 13:06:32 diablo-x570aorusmaster kernel: PM: Possible incorrect RTC due to pm_trace, please use 'ntpdate' or 'rdate' to reset it.
Jul 24 13:06:32 diablo-x570aorusmaster kernel: PM: suspend exit
In my own situation it wasn’t on a PM line directly - a search for PM: with less got me to an approximate place, and from there to read.
When I replicate I find as follows - of course this directly matches only to the issue I observed, if it’s not in your case a user space cause it could indicate differently, but I hope it’ll provide a lead.
Jul 25 09:42:30 yourhostname systemd[1]: Starting Suspend…
Jul 25 09:42:30 yourhostname systemd-sleep[1643546]: Suspending system…
Jul 25 09:42:30 yourhostname kernel: [101974.899986] PM: suspend entry (deep)
Jul 25 09:42:50 yourhostname kernel: [101975.190968] Filesystems sync: 0.290 seconds
Jul 25 09:42:50 yourhostname kernel: [101975.191307] Freezing user space processes …
Jul 25 09:42:50 yourhostname kernel: [101995.197853] Freezing of tasks failed after 20.006 seconds (1 tasks refusing to freeze, wq_busy=0):
Jul 25 09:42:50 yourhostname kernel: [101995.197883] task:fstrim state:D stack: 0 pid:1641223 ppid:1641222 flags:0x00004004