You are not logged in.
"module_blacklist=snd_hda_scodec_cs35l41"?
Sorry, don't know how I've missed that. Blacklisted.
This killed the sound on the in-built speakers, will report on the freezes.
Offline
Blacklisting the module didn't help, still got a freeze on a wake-up from suspend.
Offline
Any pattern to those freezes? Do you have to sleep the system for longer?
Do you get those freezes when suspending the system from a KDE/X11 (or even better: icewm or so) session?
Can you trigger them when suspending from the multi-user.target (no GUI running, see the 2nd link below)?
Offline
The only pattern I've noticed is that it usually occurs after longer suspends, but not always. So, inconsistent.
I'm suspending it from a KDE/Wayland session. Didn't try anything else. I have a suspicion that it's a SDDM/Kwin thing, but that's just a guess.
I didn't try to trigger them without GUI, but since I can't reliably reproduce them (repeatedly suspending and waking up doesn't work), and my laptop is my primary means of work, I can't leave it for 15-20h without GUI for the sake of testing.
Offline
Cut out SDDM and try Plasma/X11 then - either it's a red herring or we at least know where to look next.
For Plasma/X11, just a hunch: try to suspend the compositor before sleeping (SHIFT+Alt+F12 toggles it, this won't work on wayland)
Offline
Switching to Plasma/X11 didn't help. Next I'll try to change SDDM for some other display manager, I suppose.
Offline
Did you
try to suspend the compositor before sleeping
Offline
Did you
seth wrote:try to suspend the compositor before sleeping
Oh, forgot that one, actually. Will try it too.
Edit: for some reason alt+shift+F12 doesn't work, despite being enabled.
Last edited by Gricenchos (2025-05-07 08:32:18)
Offline
"Doesn't work" how?
qdbus6 org.kde.KWin /KWin supportInformation does is suggest that this is X11 (because no, won't work on wayland) and does the state of the compositor not change?
In doubt just try an openbox or icewm session…
Offline
"Doesn't work" how?
qdbus6 org.kde.KWin /KWin supportInformationdoes is suggest that this is X11 (because no, won't work on wayland) and does the state of the compositor not change?
In doubt just try an openbox or icewm session…
Yes, it's X11. The state of the compositor doesn't change whatsoever. I will try openbox, though I'm not too used to wms. But worth a try, I suppose.
Also I've noticed that the freezes only appear after long suspends, usually when I leave the laptop suspended overnight. Not sure if this has anything to do with anything.
Offline
only appear after long suspends, usually when I leave the laptop suspended overnight.
Sounds like VRAM decay…
The state of the compositor doesn't change whatsoever.
So the shortcut isn't used/enabled/confgured?
There's probably also a checkbox in systemsettings to toggle this.
Offline
only appear after long suspends, usually when I leave the laptop suspended overnight.
Sounds like VRAM decay…
The state of the compositor doesn't change whatsoever.
So the shortcut isn't used/enabled/confgured?
There's probably also a checkbox in systemsettings to toggle this.
The shortcut is enabled in the settings. But pressing it has no effect.
VRAM decay? Is there something to do about that or is that a purely hardware issue?
Offline
It does *not* change the "compositing: active" line in the kwin supportInformation??
qdbus6 org.kde.KWin /KWin supportInformation
# press SHIFT+Alt+F12
qdbus6 org.kde.KWin /KWin supportInformationhttps://wiki.archlinux.org/title/NVIDIA … er_suspend - but that's also enabled by default.
Offline
It does *not* change the "compositing: active" line in the kwin supportInformation??
qdbus6 org.kde.KWin /KWin supportInformation # press SHIFT+Alt+F12 qdbus6 org.kde.KWin /KWin supportInformationhttps://wiki.archlinux.org/title/NVIDIA … er_suspend - but that's also enabled by default.
Yes, it still says "compositing is active" after the button press, despite it definitely being enabled in the settings.
I'll look into the link, though I have an AMD video card, not Nvidia.
Offline
Sorry - in that case the nvidia wiki is irrelevant, nothing of that will work w/ amdgpu.
Does the long sleep end up becoming hibernation?
Offline
Well, switching to openbox before suspending didn't help. Neither did switching to discrete graphics from switchable. The freeze happened a few seconds after I opened the lid and started entering the password.
I have a slight suspicion that it doesn't have anything to do with suspend at all. When I was running memtest overnight, it also froze at some point. Maybe it's worth to leave the laptop active for the night and see if it still freezes?
Offline
Sorry - in that case the nvidia wiki is irrelevant, nothing of that will work w/ amdgpu.
Does the long sleep end up becoming hibernation?
Sorry for the late reply. Yes, it does.
Offline
hibernation like in S4, https://wiki.archlinux.org/title/Power_ … ibernation ?
When I was running memtest overnight, it also froze at some point.
Does not allow for https://bbs.archlinux.org/viewtopic.php … 1#p2239211 - does it?
And I wonder whether that somehow resets after the sleep => https://wiki.archlinux.org/title/Power_ … stem-sleep
Offline
And now I've got a kernel panic again; that didn't happen for a while. Report:
Offline
hibernation like in S4, https://wiki.archlinux.org/title/Power_ … ibernation ?
When I was running memtest overnight, it also froze at some point.
Does not allow for https://bbs.archlinux.org/viewtopic.php … 1#p2239211 - does it?
And I wonder whether that somehow resets after the sleep => https://wiki.archlinux.org/title/Power_ … stem-sleep
Okay, now that's interesting: it says that to hibernate you have to create a swap partition. But when making this install, I've created none: I don't have a HDD, and I've read that swap shouldn't be created on SSD at all. Might that be the case?
It's still odd that the problem only occurred after a year of usage. Maybe with the growing size of the / partition the system now can't allocate enough resources for that? Just shooting in the dark, tbh.
Offline
The system will not hibernate if you didn't configure it for that - you're most likely not hibernating.
It's still odd that the problem only occurred after a year of usage.
No, the ryzen problem is an aggrevating condition; it typically gets worse over time.
30691.479190] refill_stock+0x44/0xb0
[30691.479194] obj_cgroup_uncharge_pages+0x5d/0xf0
[30691.479198] __memcg_slab_free_hook+0xb4/0x140
[30691.479201] kmem_cache_free+0x46f/0x4d0Though I'd say the issue is currently around your RAM - you've STR issues and memtest86+ freezes and the latest panic is memory related as well.
Are you currently back to
However, it came to me that I have mismatched memory modules: one 8gb Samsung and a 16gb Kingston. Might this be the issue?
What if you remove the 16gb DIMM (suffer from 8GB for a while and then test the old 8GB DIMM again)
Offline
The system will not hibernate if you didn't configure it for that - you're most likely not hibernating.
It's still odd that the problem only occurred after a year of usage.
No, the ryzen problem is an aggrevating condition; it typically gets worse over time.
30691.479190] refill_stock+0x44/0xb0 [30691.479194] obj_cgroup_uncharge_pages+0x5d/0xf0 [30691.479198] __memcg_slab_free_hook+0xb4/0x140 [30691.479201] kmem_cache_free+0x46f/0x4d0Though I'd say the issue is currently around your RAM - you've STR issues and memtest86+ freezes and the latest panic is memory related as well.
Are you currently back toHowever, it came to me that I have mismatched memory modules: one 8gb Samsung and a 16gb Kingston. Might this be the issue?
What if you remove the 16gb DIMM (suffer from 8GB for a while and then test the old 8GB DIMM again)
Yeah, I've returned the 16bg DIMM, so I'm running 24 total right now. Are you suggesting to completely remove the non-soldered DIMM and test it like that? I will try it tomorrow.
Edit: I should've mentioned, that memtest86+ didn't find any errors when I ran it.
Last edited by Gricenchos (2025-05-10 22:02:51)
Offline
Are you suggesting to completely remove the non-soldered DIMM and test it like that?
Yup. Still along reducing PBO with the curve control service.
I should've mentioned, that memtest86+ didn't find any errors when I ran it.
But it ended up freezing?
Offline
Yup. Still along reducing PBO with the curve control service.
Will do it today and report on the results.
But it ended up freezing?
Yes, at some point at night. I don't remember how many passes it made, but it was a substantial number, I think. No errors, but completely frozen laptop in the morning.
Offline
Completely removing one of the DIMMs didn't help, still got a freeze after a night of inactivity.
Offline