wake up from sleep ends in rebooting.

ThePeach · 2024-11-27 16:49:15

Hello all, super odd situation: I'm on a Framework Laptop 13" with AMD CPU I recently swapped from Intel i7 and I'm finding myself some odd stuff here and there (and to be honest I start to think I should have reinstalled everything from scratch.. but before I even try to do anything like that, I was wondering if you could help me debug this (as I'm about to head off on another continent).

Long story short, I put the laptop to sleep, go to wake it up after a good while (this is not consistent, seems like it happens more frequently the more I leave it there) and it just reboots.

When I look at the kernel logs, this is what I see:

-1 log

Nov 27 16:36:59 frametree kernel: PM: suspend entry (s2idle)
Nov 27 16:36:59 frametree systemd-sleep[9165]: Performing sleep operation 'suspend'...
Nov 27 16:36:59 frametree systemd-sleep[9165]: Successfully froze unit 'user.slice'.
Nov 27 16:36:59 frametree wpa_supplicant[852]: wlp1s0: CTRL-EVENT-DSCP-POLICY clear_all
Nov 27 16:36:59 frametree systemd[1]: Starting System Suspend...
Nov 27 16:36:59 frametree systemd[1]: Reached target Sleep.
....

current log start:

....
Nov 27 17:15:12 frametree kernel: BIOS-provided physical RAM map:
Nov 27 17:15:12 frametree kernel: Command line: BOOT_IMAGE=/vmlinuz-linux root=UUID=e1f28acf-a705-425d-8539-480277a46eb8 rw loglevel=3 quiet
Nov 27 17:15:12 frametree kernel: Linux version 6.12.1-arch1-1 (linux@archlinux) (gcc (GCC) 14.2.1 20240910, GNU ld (GNU Binutils) 2.43.0) #1 SMP PRE>

a successful wake up from suspend would look like this:

...
Nov 27 17:46:43 frametree kernel: [drm] Skip DMUB HPD IRQ callback in suspend/resume
Nov 27 17:46:43 frametree kernel: [drm] Skip DMUB HPD IRQ callback in suspend/resume
Nov 27 17:46:43 frametree kernel: amdgpu 0000:c1:00.0: amdgpu: SMU is resumed successfully!
Nov 27 17:46:43 frametree kernel: nvme nvme0: 16/0/0 default/read/poll queues
Nov 27 17:46:43 frametree kernel: amdgpu 0000:c1:00.0: amdgpu: SMU is resuming...
Nov 27 17:46:43 frametree kernel: [drm] PCIE GART of 512M enabled (table at 0x000000801FD00000).
Nov 27 17:46:43 frametree kernel: ACPI: EC: interrupt unblocked
Nov 27 17:46:43 frametree kernel: ACPI: EC: interrupt blocked
Nov 27 17:46:43 frametree kernel: pcieport 0000:00:08.3: quirk: disabling D3cold for suspend
Nov 27 17:46:43 frametree kernel: queueing ieee80211 work while going to suspend
Nov 27 17:46:43 frametree kernel: atkbd serio0: Disabling IRQ1 wakeup source to avoid platform firmware bug
Nov 27 17:46:43 frametree kernel: printk: Suspending console(s) (use no_console_suspend to debug)
Nov 27 17:46:43 frametree kernel: Freezing remaining freezable tasks completed (elapsed 0.001 seconds)
Nov 27 17:46:43 frametree kernel: Freezing remaining freezable tasks
Nov 27 17:46:43 frametree kernel: OOM killer disabled.
Nov 27 17:46:43 frametree kernel: Freezing user space processes completed (elapsed 0.002 seconds)
Nov 27 17:46:43 frametree kernel: Freezing user space processes
Nov 27 17:46:31 frametree kernel: Filesystems sync: 0.002 seconds
Nov 27 17:46:31 frametree kernel: PM: suspend entry (s2idle)
...

I have no idea where to start on this... any hints?

seth · 2024-11-28 07:51:57

1st snippet: does the journal for the boot end there?
3rd snippet actually ends where the 1st one starts, so we've no overlap to compare.
2nd snippet: you're looking for MCE errors which would be logged somewhat later on

* You're using s2idle, can you try s3?
https://wiki.archlinux.org/title/Power_ … end_method
* Is there a parallel windows installation?
* Is the amdgpu a dedicated device or the new APU?

ThePeach · 2024-12-03 14:27:57

seth wrote:

1st snippet: does the journal for the boot end there?
3rd snippet actually ends where the 1st one starts, so we've no overlap to compare.
2nd snippet: you're looking for MCE errors which would be logged somewhat later on

ok will check later

* You're using s2idle, can you try s3?

only method available is s2idle:

$  cat /sys/power/mem_sleep
[s2idle]

so I'd rather try to see if I can fix this, I feel like there some difference conf between intel and AMD boards that is causing this...

* Is there a parallel windows installation?

you mean DE? now I have both Gnome and KDE installed, if that's what you're asking, tried to move away from Gnome, without uninstalling it completely...

* Is the amdgpu a dedicated device or the new APU?

came all together with the same board.

seth · 2024-12-03 15:59:37

you mean DE?

No, like MS Windows - a second operating system: https://wiki.archlinux.org/title/Dual_boot

came all together with the same board.

So you swapped the entire board, new BIOS, new bus system, new … basically everything relevant.
That's gonna cause differences, yes.

ThePeach · 2024-12-03 18:47:39

seth wrote:

you mean DE?
No, like MS Windows - a second operating system: https://wiki.archlinux.org/title/Dual_boot

nope, just Linux

So you swapped the entire board, new BIOS, new bus system, new … basically everything relevant.
That's gonna cause differences, yes.

that I seem to have noticed. Currently the only problem I'm having is with the suspend, the rest of the machine behaves really well, connection, buses, devices and so on. So even with this little hitch, I'm quite impressed.

I'll try to dig further with this issue and see if I can get some more info on what could cause this.
Thanks for any help so far

Last edited by ThePeach (2024-12-04 02:34:58)

seth · 2024-12-03 21:47:07

If you want a more informed comment, we'll have to see relevant logs.
Generally test the LTS kernel, look out for BIOS updates and test

amdgpu.aspm=0 amdgpu.runpm=0 amdgpu.bapm=0 pcie_aspm=off nvme_core.default_ps_max_latency_us=0 iommu=soft processor.max_cstate=5 amdgpu.ppfeaturemask=0xffcff8

The last kernel parameter might be equivalent to amdgpu.dpm=0 which has a habit of completely breaking the boot, so if the system doesn't boot w/ those, remove that one first.
The stuff disables a lot of power saving mechanisms, so your system will drain the battery faster this way - it's just a blind guess on what might cause the reboots itfp.

Also see the entire https://wiki.archlinux.org/title/Ryzen#Troubleshooting section.

Arch Linux

#1 2024-11-27 16:49:15

wake up from sleep ends in rebooting.

#2 2024-11-28 07:51:57

Re: wake up from sleep ends in rebooting.

#3 2024-12-03 14:27:57

Re: wake up from sleep ends in rebooting.

#4 2024-12-03 15:59:37

Re: wake up from sleep ends in rebooting.

#5 2024-12-03 18:47:39

Re: wake up from sleep ends in rebooting.

#6 2024-12-03 21:47:07

Re: wake up from sleep ends in rebooting.

Board footer