[SOLVED] Everything freezes on poweroff or restart

TanX-009 · 2024-07-23 15:55:35

Everything freezes somewhere in between the waterfall

It doesn't occur always.
It has been happening since 2-3 weeks now

Sometimes there is even a kernel panic
I'll share the logs as soon as another occurs(kernel panic)

last poweroff
http://0x0.st/XpWu.txt

I'm dual booted with windows and fast start is turned off.

Specs: Ryzen 7 4800h
RTX 3050
16gb 3200mhz
512gb Intel SSD + 256gb Samsung SSD

Last edited by TanX-009 (2024-09-10 08:20:35)

WayNKG · 2024-07-23 16:59:29

Full on freezes and doesn't shutdown? Or like, says something about stop jobs or something? Cause the 2nd one happened to me a bunch.

seth · 2024-07-23 22:01:22

THe big red flöag are the nvidiaz drivers, https://bbs.archlinux.org/viewtopic.php … 6#p2182836
Test the 535xx ones.

TanX-009 · 2024-07-24 06:09:33

WayNKG wrote:

Full on freezes and doesn't shutdown?

Full on freezes.
The stop jobs are understandable and recognisable.

seth wrote:

Test the 535xx ones.

I'll try those.

TanX-009 · 2024-07-24 07:29:55

how do i properly downgrade nvidia?
downgrading only nvidia-dkms, nvidia-utils
causes multiple problems

last time i tried downgrading it nothing worked
this time after downgrading nvidia-dkms, nvidia-utils, opencl-nvidia to 535.98 it wont boot into hyprland
and there are errors with egl
Crash logs

Last edited by TanX-009 (2024-07-24 07:35:11)

TanX-009 · 2024-07-24 12:44:44

Another crash but now it crashed while installing package at the stage of reloading system ...
now I've completely bricked
Reinstalling now got these logs from otg installation
http://0x0.st/Xp3h.txt

This happened once before (is this related?)
and once before that but that was due to memory frequency mismatch 2400mhz and 3200mhz
I exchanged it with my friend's now 3200mhz both
and checked using memtest passed after swapping

Last edited by TanX-009 (2024-07-24 12:55:50)

seth · 2024-07-24 18:34:41

Another crash but now it crashed while installing package at the stage of reloading system ...

This is a common issue w/ the 550xx and probably 555xx drivers - iff they corrupt the kernel memry this seems to affect cgroups and induce a crash in system when the deamon tries to reload what unfortunately happens during updates.

wrt the 535xx drivers, https://aur.archlinux.org/packages/nvidia-535xx-dkms is at 535.183 - do NOT just pull some packages out of the ALA here!

TanX-009 · 2024-07-25 03:32:04

Ohhh! thats why when i downloaded older package from aur which was incompatible with current kernel and hence not work.
Well didn't know that.

seth wrote:

wrt the 535xx drivers, https://aur.archlinux.org/packages/nvidia-535xx-dkms is at 535.183 - do NOT just pull some packages out of the ALA here!

will try that

Edit
tried nvidia-535xx-dkms but
using linux-zen btw

==> ERROR: module not found: 'nvidia'
==> ERROR: module not found: 'nvidia_modeset'
==> ERROR: module not found: 'nvidia_uvm'
==> ERROR: module not found: 'nvidia_drm'

Last edited by TanX-009 (2024-07-25 03:47:17)

seth · 2024-07-25 06:19:50

dkms status
pacman -Qs kernel

If you installed some dkms package from the ala, chances are you previously didn't build a module either?
What's the otput of "dkms autoinstall"?

TanX-009 · 2024-07-25 13:43:02

dkms status

http://0x0.st/XfXf.txt

pacman -Qs kernel

http://0x0.st/XfXV.txt

dkms autoinstall

http://0x0.st/XfXv.txt

seth · 2024-07-25 17:22:11

You don't really have to paste onliners…

nvidia/535.183.01: added

So it's not built (would say "installed")

local/linux-zen 6.10.zen1-2
    The Linux ZEN kernel and modules
local/linux-zen-headers 6.10.zen1-2
    Headers and scripts for building modules for the Linux ZEN kernel

kernel and matching headers are installed…

But the autoinstall contains almost nothing (exits w/ error 2)
Did you forget to redirect stderr?
There's typically gonna be a make log and dkms will tell yo to look there for details.

Edit: https://bbs.archlinux.org/viewtopic.php … 8#p2186198
It's b/c of 6.10…

Last edited by seth (2024-07-25 17:25:36)

TanX-009 · 2024-07-26 04:57:50

seth wrote:

You don't really have to paste onliners…

Ok.(Its easy on android when in tty)
So wait untill another update for linux is released
Or
Downgrade linux?
That didn't go well last time
How to do that properly?

Last edited by TanX-009 (2024-07-26 05:00:17)

seth · 2024-07-26 05:42:06

There's a patch to compile w/ 6.10 linked on the 535xx dkms AUR, https://gitlab.com/herecura/packages/nv … 6.10.patch
Alternatively you can downgrade linux-zen and linux-zen-headers to the last 6.9 version from the ALA, then build the 535xx dkms kernel from the AUR (which you already have) against that (should™ even happen automatically when installing the older kernel) - that's not like trying to use an older version of nvidia-utils which was like userspace-ABI incompatible.

TanX-009 · 2024-07-26 10:43:54

downgrading linux and headers to 6.9.10 there were no errors in mkinitcpio and dkms autoinstall
but hyprland didn't start and there weren't any errors or crashreport

Last edited by TanX-009 (2024-07-26 10:44:14)

seth · 2024-07-26 17:47:21

but hyprland didn't start and there weren't any errors or crashreport

What happens when you boot the multi-user.target (2nd linke below), login and run "Hyprland"?
https://wiki.archlinux.org/title/Hyprland#Starting

TanX-009 · 2024-07-30 18:29:06

With linux-zen 6.10.2 and nvidia-535xx-dkms and nvidia-535xx-utils
Same as downgrading kernel
Hyprland doesn't launch and no errors or crashreport

Same behaviour with multi-user.target

seth · 2024-07-30 20:25:30

Hyprland doesn't launch and no errors or crashreport

That's not a thing, there's gonna be some return code or error message from the shell or *something*, entering something into a terminal will never just cause "nothing"

Hyprland; echo $?

To see coredumps, https://wiki.archlinux.org/title/Core_d … _core_dump
Check the system journal "sudo journalctl -b" and try to move away your ~/.config/hypr/hyprland.conf

Generally please post your complete system journal for the boot:

sudo journalctl -b | curl -F 'file=@-' 0x0.st

TanX-009 · 2024-07-31 05:50:57

seth wrote:

That's not a thing, there's gonna be some return code or error message from the shell or *something*, entering something into a terminal will never just cause "nothing"

On launch usually the last line says about the crashreport
but it didn't

Hyprland; echo $?

showed that line

crashreport
journalctl

While nvidia-535xx-dkms... is installed I tried switching to Integrated using supergfxctl
but was unsuccessful

Last edited by TanX-009 (2024-07-31 05:52:11)

seth · 2024-07-31 06:34:30

Jul 31 10:58:24 tanx supergfxd[711]: DEBUG: Did modprobe for driver nvidia_modeset

but the module doesn't show up.
Enable https://wiki.archlinux.org/title/NVIDIA … de_setting - use the "nvidia_drm.modeset=1" kernel parameter (modprobe.conf won't do!)

Also

Jul 31 10:59:02 tanx systemd-coredump[1078]: Process 1052 (Hyprland) of user 1000 dumped core.
                                             
                                             Stack trace of thread 1052:
                                             #0  0x000072d6e30773f4 n/a (libc.so.6 + 0x963f4)
                                             #1  0x000072d6e301e120 raise (libc.so.6 + 0x3d120)
                                             #2  0x000072d6e30054c3 abort (libc.so.6 + 0x244c3)
                                             #3  0x00006144b2b12577 _Z25handleUnrecoverableSignali (Hyprland + 0x145577)
                                             #4  0x000072d6e301e1d0 n/a (libc.so.6 + 0x3d1d0)
                                             #5  0x000072d6b5c154ab n/a (nvidia-drm_gbm.so + 0x154ab)
                                             #6  0x000072d6b5c15778 n/a (nvidia-drm_gbm.so + 0x15778)
                                             #7  0x000072d6e363b1c9 n/a (libgbm.so.1 + 0x31c9)
                                             #8  0x000072d6e363b37d n/a (libgbm.so.1 + 0x337d)
                                             #9  0x000072d6e363b480 gbm_create_device (libgbm.so.1 + 0x3480)
                                             #10 0x00006144b2e190c5 wlr_gbm_allocator_create (Hyprland + 0x44c0c5)
                                             #11 0x00006144b2e1938c allocator_autocreate_with_drm_fd (Hyprland + 0x44c38c)
                                             #12 0x00006144b2e2745f init_drm_renderer (Hyprland + 0x45a45f)
                                             #13 0x00006144b2e208f9 wlr_drm_backend_create (Hyprland + 0x4538f9)
                                             #14 0x00006144b2e20b90 n/a (Hyprland + 0x453b90)
                                             #15 0x00006144b2e21239 wlr_backend_autocreate (Hyprland + 0x454239)
                                             #16 0x00006144b2b1a13a _ZN11CCompositor10initServerEv (Hyprland + 0x14d13a)
                                             #17 0x00006144b2ad6123 main (Hyprland + 0x109123)
                                             #18 0x000072d6e3006e08 n/a (libc.so.6 + 0x25e08)
                                             #19 0x000072d6e3006ecc __libc_start_main (libc.so.6 + 0x25ecc)
                                             #20 0x00006144b2b0f255 _start (Hyprland + 0x142255)
                                             
                                             Stack trace of thread 1060:
                                             #0  0x000072d6e3071a19 n/a (libc.so.6 + 0x90a19)
                                             #1  0x000072d6e3074479 pthread_cond_wait (libc.so.6 + 0x93479)
                                             #2  0x000072d6d049d3ee n/a (radeonsi_dri.so + 0x9d3ee)
                                             #3  0x000072d6d047b19c n/a (radeonsi_dri.so + 0x7b19c)
                                             #4  0x000072d6d049d31d n/a (radeonsi_dri.so + 0x9d31d)
                                             #5  0x000072d6e307539d n/a (libc.so.6 + 0x9439d)
                                             #6  0x000072d6e30fa49c n/a (libc.so.6 + 0x11949c)

both backtraces go through the nvidia GBM, but on the radeon dri

printenv

and also your hyprland config.

Also, probably just remove supergfxwhatever, you've an Ampere chip capable of and defaulting to https://wiki.archlinux.org/title/PRIME# … Management

TanX-009 · 2024-08-03 06:23:54

was using supergfxd along with asusctl
removed supergfxd

seth wrote:

Enable https://wiki.archlinux.org/title/NVIDIA … de_setting - use the "nvidia_drm.modeset=1" kernel parameter (modprobe.conf won't do!)

cat /sys/module/nvidia_drm/parameters/modeset

shows

when set from (as instructed in hyprland wiki)

# /etc/modprobe.d/nvidia.conf
blacklist nouveau
alias nouveau off

options nvidia_drm modeset=1 fbdev=1

seth wrote:

both backtraces go through the nvidia GBM, but on the radeon dri

Ahhh!, now I remember.
I had this problem with those drivers (github issue) where

env = GBM_BACKEND,nvidia-drm

caused crash in hyprland

commenting that fixed the hyprland launch

Now for the random crashes I will post as soon as it happens, if it happens in the following week.
or mark as solved

Last edited by TanX-009 (2024-08-03 06:26:32)

seth · 2024-08-03 06:35:18

Ftr, the point of the kernel parameter would have been to get rid of the simpledrm device, but the cause was likely supergfxd

TanX-009 · 2024-09-10 08:20:11

No crashes after downgrading to nvidia-535xx-dkms
Marking as solved
Though, should i updated drivers ever?
If I should, then when?

seth · 2024-09-10 13:15:37

nvidia has claimed to have a fix in an (at the time not officially released) 550xx driver (that wasn't ported to 555xx not it seems 560xx - there're still claims of it crashing)
I assume they'll fix it "eventually", but when that is remains everyones guess.

https://bbs.archlinux.org/viewtopic.php … 1#p2187361 - I'd probably subscribe to that thread to get updates on the situation.

Arch Linux

#1 2024-07-23 15:55:35

[SOLVED] Everything freezes on poweroff or restart

#2 2024-07-23 16:59:29

Re: [SOLVED] Everything freezes on poweroff or restart

#3 2024-07-23 22:01:22

Re: [SOLVED] Everything freezes on poweroff or restart

#4 2024-07-24 06:09:33

Re: [SOLVED] Everything freezes on poweroff or restart

#5 2024-07-24 07:29:55

Re: [SOLVED] Everything freezes on poweroff or restart

#6 2024-07-24 12:44:44

Re: [SOLVED] Everything freezes on poweroff or restart

#7 2024-07-24 18:34:41

Re: [SOLVED] Everything freezes on poweroff or restart

#8 2024-07-25 03:32:04

Re: [SOLVED] Everything freezes on poweroff or restart

#9 2024-07-25 06:19:50

Re: [SOLVED] Everything freezes on poweroff or restart

#10 2024-07-25 13:43:02

Re: [SOLVED] Everything freezes on poweroff or restart

#11 2024-07-25 17:22:11

Re: [SOLVED] Everything freezes on poweroff or restart

#12 2024-07-26 04:57:50

Re: [SOLVED] Everything freezes on poweroff or restart

#13 2024-07-26 05:42:06

Re: [SOLVED] Everything freezes on poweroff or restart

#14 2024-07-26 10:43:54

Re: [SOLVED] Everything freezes on poweroff or restart

#15 2024-07-26 17:47:21

Re: [SOLVED] Everything freezes on poweroff or restart

#16 2024-07-30 18:29:06

Re: [SOLVED] Everything freezes on poweroff or restart

#17 2024-07-30 20:25:30

Re: [SOLVED] Everything freezes on poweroff or restart

#18 2024-07-31 05:50:57

Re: [SOLVED] Everything freezes on poweroff or restart

#19 2024-07-31 06:34:30

Re: [SOLVED] Everything freezes on poweroff or restart

#20 2024-08-03 06:23:54

Re: [SOLVED] Everything freezes on poweroff or restart

#21 2024-08-03 06:35:18

Re: [SOLVED] Everything freezes on poweroff or restart

#22 2024-09-10 08:20:11

Re: [SOLVED] Everything freezes on poweroff or restart

#23 2024-09-10 13:15:37

Re: [SOLVED] Everything freezes on poweroff or restart

Board footer