You are not logged in.
So I just finished upgrading my PC and booted up my existing Arch Linux install (I wanted to see how well that would work) and I'm faced with a weird new issue: When attempting to suspend in Arch (from KDE Plasma) the entire system seems to fail to suspend and instead reboots. The screen turns off and after a few seconds the PC completely powers off, then turns back on and completely reboots.
Specs:
Intel Core i7-4790
Asus H97M-E/CSM mATX motherboard
MSI Mech 2X Radeon RX 6600
Google turns up nothing helpful as usual.
Last edited by EJSnow (Today 04:10:56)
I hate the rabbit hole I have fallen down but the sunk-cost fallacy, pacman, and the AUR prevent me from picking up and moving to Fedora or something.
Offline
Google usually turns up what you asked for ![]()
Do you have a journal of the previous boot after the undesired reboot?
Please post your complete system journal for the boot:
sudo journalctl -b -1 | curl -s -H "Accept: application/json, */*" --upload-file - 'https://paste.c-net.org/'Is there a parallel windows installation?
Offline
Here's the journal: https://paste.c-net.org/GonerRepeats
I do not have Windows installed right now, it's just Arch on my SSD. (And even if I did, I would have fast start disabled.)
(Edit: re-uploaded the journal after attempting to suspend)
Last edited by EJSnow (2026-06-26 17:38:11)
I hate the rabbit hole I have fallen down but the sunk-cost fallacy, pacman, and the AUR prevent me from picking up and moving to Fedora or something.
Offline
Dies out of S3 (ifff it wakes up or the S3 fails otherwise that's no longer logged)
Core(TM) i7-4790
Jun 26 13:35:59 RazorCrest systemd-sleep[31417]: /etc/systemd/sleep.conf.d/hybrid-sleep.conf:1: Assignment outside of section. Ignoring.What does that file look like?
Does s2idle behave better?
https://wiki.archlinux.org/title/Power_ … end_method
Offline
Dies out of S3 (ifff it wakes up or the S3 fails otherwise that's no longer logged)
Core(TM) i7-4790Jun 26 13:35:59 RazorCrest systemd-sleep[31417]: /etc/systemd/sleep.conf.d/hybrid-sleep.conf:1: Assignment outside of section. Ignoring.What does that file look like?
ejsnow in ~
❯ cat /etc/systemd/sleep.conf.d/hybrid-sleep.conf 02:22:57 PM
hybrid-sleepDoes s2idle behave better?
https://wiki.archlinux.org/title/Power_ … end_method
With the suspend method set to s2idle: Now when suspending the screen turns off, but the fans continue to spin. Pressing any key on the keyboard, moving/clicking the mouse, a short press of the power button all do nothing; the system is stuck in limbo and the only way out is to hard-reset it by holding the power button for 10 seconds.
I would hardly call it an improvement.
I hate the rabbit hole I have fallen down but the sunk-cost fallacy, pacman, and the AUR prevent me from picking up and moving to Fedora or something.
Offline
What is that file?
Remove that.
Try to disable SYSTEMD_SLEEP_FREEZE_USER_SESSIONS=false, https://bbs.archlinux.org/viewtopic.php?id=296954
Can you suspend from the console?
What if you stop/disable the foreground-booster thing?
(s3 and s2idle probably behave the same, leaving the acpi state aside that causes a power loss in s3)
Offline
What is that file?
Remove that.
Removed. Nothing changes. (It may be related to the one time I tried hybrid sleep a while ago on my old PC?)
As for SYSTEMD_SLEEP_FREEZE_USER_SESSIONS, I've done the following:
$ cat /etc/systemd/system/systemd-suspend.service.d/override.conf
[Service]
Environment="SYSTEMD_SLEEP_FREEZE_USER_SESSIONS=false"Not sure if I need to apply this for anything else? This also doesn't help, however.
Can you suspend from the console?
No. It still does the exact same thing when running systemctl suspend on a console-only boot (disabled plasmalogin.service). Journal for console-only boot: https://paste.c-net.org/DoubtingDefeat
I'll look at what happens with the foreground-booster stuff removed. It's supposed to help game performance in VRAM-limited scenarios (mostly with low amounts of VRAM like 4GB or less) but now that I have an RX 6600 (with 8GB of VRAM) it probably isn't that helpful anymore.
I hate the rabbit hole I have fallen down but the sunk-cost fallacy, pacman, and the AUR prevent me from picking up and moving to Fedora or something.
Offline
Jun 26 15:17:45 RazorCrest systemd-sleep[1229]: User sessions remain unfrozen on explicit request ($SYSTEMD_SLEEP_FREEZE_USER_SESSIONS=0).So not that.
lact kicks in when the system sleeps …
Offline
I did see that. It appears it's just reloading/reapplying configs after suspend/resume. I was previously using it to undervolt my old GPU (an RX 6400) but for now I'm not using it (though I plan to undervolt my GPU at some point). I could see what happens if I disable it...
I hate the rabbit hole I have fallen down but the sunk-cost fallacy, pacman, and the AUR prevent me from picking up and moving to Fedora or something.
Offline
Another thing that may or may not be relevant is that my PC doesn't want to stay powered off when I shut it down. I have to turn off the power supply, otherwise it turns itself back on after shutting down. I did have this problem on my old PC but that turned out to be wake on LAN being broken and it only happened after shutting down from Linux specifically. On my new PC it happens when shutting down from the BIOS as well. The BIOS settings are a little confusing but wake-on-LAN appears to be disabled on my PC, so it has to be something else. I'm starting to think this may be a hardware issue and no fault of Linux's. Although I haven't tried Windows (or using the acpi_osi kernel parameter to make the board think I'm using Windows).
Edit: I've both disabled LACT and AMD Overdrive and added the following to my kernel parameters:
acpi_osi=! acpi_osi="Windows 2022"This still doesn't fix it. That leaves me pretty convinced I have a hardware issue.
Last edited by EJSnow (Yesterday 03:19:54)
I hate the rabbit hole I have fallen down but the sunk-cost fallacy, pacman, and the AUR prevent me from picking up and moving to Fedora or something.
Offline
That leaves me pretty convinced I have a hardware issue.
my PC doesn't want to stay powered off when I shut it down. I have to turn off the power supply, otherwise it turns itself back on after shutting down.
Jun 26 15:16:50 archlinux kernel: DMI: ASUS All Series/H97M-E, BIOS 2702 03/28/2016just finished upgrading my PC and booted up my existing Arch Linux install […] the entire system seems to fail to suspend and instead reboots
What exactly was upgraded there? Did the PC do the same reboot-on-poweroff dance before that?
Offline
So the upgrade was from a Dell optiplex 7020 sff (Intel Q87 chipset) with the same CPU, ram and an RX 6400 gpu to my current system (the main upgrade was the RX 6600 and I needed to upgrade the power supply, case, and motherboard as well to handle it). The old system did have a similar reboot-on-poweroff issue which I posted about here: https://bbs.archlinux.org/viewtopic.php?id=304996 which I solved by disabling Wake-on-LAN. The thing is, my new-not-new motherboard disables Wake-on-LAN by default.
Anyways, I had a power blip due to a thunderstorm in my area this evening. I did shut down my PC beforehand, but when I turned it back on later it defaulted the BIOS settings so I had to re-set them. I'm not even sure what setting exactly was responsible (I changed a lot of things) but oddly enough, after booting up Arch, I tried to suspend it and it... suspended. I suspect a particular BIOS setting I changed may have been responsible so I will go hunt through my BIOS settings again...
Edit: Hunted through the BIOS settings. None of the options I thought would have an effect did anything.
I LOVE working with PCs. I really really love working with PCs.
I think the power blip may have somehow caused a complete CMOS reset (although the date/time still being correct when I booted it up later, however it greeted me with an AMI screen saying basically options were defaulted go to the BIOS and reconfigure). And somehow or other that made it work???? Regardless, it seems to indicate... a dead/dying CMOS battery.
Edit #2: Seems I got a 2-for-one as I tried just shutting down my computer... and it stays off now. It doesn't randomly reboot. My best guess is somehow a CMOS reset was triggered and that kind of reset something somehow and now power management just works? (Oh and interesting fact: the BIOS has an option to toggle S4 support (hibernate) and it defaults to off although I enabled it. No, toggling it does not mess with S3 or S5 states. Weird.) Oh well seems my issue was just a classic case of PCs working Perfectly™ with Zero Flaws™ Every Time™.
Last edited by EJSnow (Today 04:10:43)
I hate the rabbit hole I have fallen down but the sunk-cost fallacy, pacman, and the AUR prevent me from picking up and moving to Fedora or something.
Offline