[Solved] nvidia drivers not detecting GPU after suspending it

dybbuk · 2025-07-24 07:05:34

after a recent nvidia update, when i my gpu wakes from susped then goes suspended again, when i try to use it it says

~$ nvidia-smi 
Unable to determine the device handle for GPU0: 0000:01:00.0: Unknown Error
No devices were found

journalctl: https://0x0.st/8nDk.txt

specs:
6.12.39-1-lts (i have tried the zen kernel, same issue)
nvidia: nvidia-dkms 575.64.03-1
GPU: nvidia rtx 2050
CPU: intel core i5-11400H
laptop: ASUS TUF Gaming F15 FX506HF_FX506HF

Last edited by dybbuk (2025-08-04 10:51:53)

seth · 2025-07-24 11:32:49

when i my gpu wakes from susped

What does that mean?
The journal covers a 80 minute boot, there's no indication of s2idle, S3 or S4 - asusd and supergfxctl are running but there#s no indication of a mode switch nor anything, really between boot and shutdown

Jul 24 01:47:19 dyb1b0k systemd[1]: Starting Hostname Service...
Jul 24 01:47:19 dyb1b0k systemd[1]: Started Hostname Service.
Jul 24 01:49:08 dyb1b0k systemd[1]: systemd-hostnamed.service: Deactivated successfully.
Jul 24 01:54:24 dyb1b0k systemd[1]: Starting Cleanup of Temporary Directories...
Jul 24 01:54:25 dyb1b0k systemd[1]: systemd-tmpfiles-clean.service: Deactivated successfully.
Jul 24 01:54:25 dyb1b0k systemd[1]: Finished Cleanup of Temporary Directories.
Jul 24 02:01:15 dyb1b0k bluetoothd[642]: src/profile.c:ext_io_disconnected() Unable to get io data for Hands-Free Voice gateway: getpeername: Transport endpoint is not connected (107)
Jul 24 02:21:34 dyb1b0k systemd[1]: Starting Hostname Service...
Jul 24 02:21:35 dyb1b0k systemd[1]: Started Hostname Service.
Jul 24 02:22:05 dyb1b0k systemd[1]: systemd-hostnamed.service: Deactivated successfully.
Jul 24 02:22:26 dyb1b0k systemd[1]: Starting Hostname Service...
Jul 24 02:22:26 dyb1b0k systemd[1]: Started Hostname Service.
Jul 24 02:22:56 dyb1b0k systemd[1]: systemd-hostnamed.service: Deactivated successfully.
Jul 24 02:50:34 dyb1b0k systemd[1]: Starting Hostname Service...
Jul 24 02:50:34 dyb1b0k systemd[1]: Started Hostname Service.
Jul 24 02:51:05 dyb1b0k systemd[1]: systemd-hostnamed.service: Deactivated successfully.
Jul 24 02:51:28 dyb1b0k systemd[1]: Starting Hostname Service...
Jul 24 02:51:28 dyb1b0k systemd[1]: Started Hostname Service.
Jul 24 02:53:24 dyb1b0k systemd[1]: systemd-hostnamed.service: Deactivated successfully.
Jul 24 02:54:05 dyb1b0k systemd[1]: Starting Hostname Service...
Jul 24 02:54:05 dyb1b0k systemd[1]: Started Hostname Service.
Jul 24 02:55:13 dyb1b0k systemd[1]: systemd-hostnamed.service: Deactivated successfully.
Jul 24 02:57:37 dyb1b0k kernel: perf: interrupt took too long (2612 > 2500), lowering kernel.perf_event_max_sample_rate to 76500
Jul 24 03:04:17 dyb1b0k systemd-logind[671]: Power key pressed short.
Jul 24 03:04:17 dyb1b0k systemd-logind[671]: Powering off...
Jul 24 03:04:17 dyb1b0k asusd[874]: [DEBUG asusd] Doing on_prepare_for_shutdown(true)
Jul 24 03:04:17 dyb1b0k polkitd[838]: Unregistered Authentication Agent for unix-session:c1 (system bus name :1.35, object path /org/kde/PolicyKit1/AuthenticationAgent, locale en_US.UTF-8) (disconnected from bus)
Jul 24 03:04:17 dyb1b0k systemd-logind[671]: System is powering down.
Jul 24 03:04:17 dyb1b0k bluetoothd[642]: Terminating
Jul 24 03:04:17 dyb1b0k systemd[1]: Stopping Session c1 of User dy...

The unprovoked hostname cycling looks a bit sketchy, did you filter the journal?

dybbuk · 2025-07-24 19:39:00

seth wrote:

What does that mean?

~$ cat /sys/bus/pci/devices/0000:01:00.0/power/control
auto

i meant when the gpu is not being used it goes into low power state or suspended (sorry about the typo)
but then when i call `nvidia-smi` it's supposed to wake up (which takes a few secs), but instead it says "No devices were found"

this is how i got the journal

# sudo journalctl -b -1 | curl -F 'file=@-' 0x0.st
http://0x0.st/8nFW.txt

dybbuk · 2025-07-24 19:50:50

i forgot to mentions something

when that issue happens, i cant shut down my laptop
only way is through REISUB then shutdown again

seth · 2025-07-24 19:52:07

Your GPU will default to https://wiki.archlinux.org/title/PRIME# … Management - does the same happen w/o asusd/supergfxctl?

Edit

Jul 24 19:23:58 dyb1b0k systemd-networkd[412]: Meta: Lost carrier
Jul 24 19:24:17 dyb1b0k systemd-logind[681]: Power key pressed short.
Jul 24 19:24:17 dyb1b0k systemd-logind[681]: Powering off...
Jul 24 19:24:17 dyb1b0k systemd-logind[681]: System is powering down.

Why can't you shut down otherwise? Can you still switch the VT (ctrl+alt+f3)

nb. that

Jul 24 19:24:17 dyb1b0k systemd[1]: ly.service: Main process exited, code=exited, status=15/n/a
Jul 24 19:24:17 dyb1b0k systemd[1]: ly.service: Failed with result 'exit-code'.

systematically breaks session integration what means you might have to "sudo shutdown" just because of that.

Last edited by seth (2025-07-24 19:55:36)

dybbuk · 2025-07-24 19:55:10

ig yeh, i can get rid of asusd if i have to

i never had that issue untill recently and i been using asusd/supergfxctl for a long time

Last edited by dybbuk (2025-07-24 20:04:25)

dybbuk · 2025-07-27 20:45:01

seth wrote:

systematically breaks session integration what means you might have to "sudo shutdown" just because of that

my main issue is not shutdown, it's the fact that the gpu don't get recognized and i have to reboot in order for it to work

seth · 2025-07-27 22:05:55

my main issue is not shutdown

I got that, I just wanted to point out that an unprivileged shutdown might fail for you (because you mentioned you cannot shutdown) just because ly trashed your session and unrelated to the main issue.

Have you been running w/o asusd/supergfxctl in the meantime and did that have any impact?

dybbuk · 2025-07-28 00:17:49

seth wrote:

did that have any impact?

no i've always been using asusd/supergfxctl and they are not the issue, it started happening recently

Edit:
shutdown/reboot even hangs with sudo

Last edited by dybbuk (2025-07-28 00:20:54)

seth · 2025-07-28 06:54:37

no i've always been using asusd/supergfxctl and they are not the issue, it started happening recently

To be very clear:

seth wrote:

Have you been running w/o asusd/supergfxctl in the meantime

yes or no?
Believing what is and is not the issue based on past behavior is a fallacy: *everything* worked in the past and now it doesn't and unless you don't know the cause, everything could be the cause.

Wrt to journal you posted, there's no nvidia related error logged at all, nor are there any bus errors.
Did you actually try to run nvidia-smi at the time?

Jul 24 19:22:56 dyb1b0k kernel: tun: Universal TUN/TAP device driver, 1.6
Jul 24 19:22:56 dyb1b0k kernel: netlink: 'verge-mihomo': attribute type 22 has an invalid length.
Jul 24 19:22:56 dyb1b0k iwd[662]: udev interface=Meta ifindex=6
Jul 24 19:22:56 dyb1b0k systemd-resolved[401]: Meta: Bus client set search domain list to: ~.
Jul 24 19:22:56 dyb1b0k systemd-resolved[401]: Using degraded feature set UDP instead of UDP+EDNS0 for DNS server 1.1.1.1.
Jul 24 19:22:56 dyb1b0k systemd-resolved[401]: Meta: Bus client set default route setting: yes
Jul 24 19:22:56 dyb1b0k systemd-resolved[401]: Meta: Bus client set DNS server list to: 198.18.0.2
Jul 24 19:23:00 dyb1b0k systemd[1]: Starting Daemon for power management...
Jul 24 19:23:00 dyb1b0k systemd[1]: Started Daemon for power management.
Jul 24 19:23:07 dyb1b0k systemd-timesyncd[371]: Timed out waiting for reply from 198.18.0.4:123 (2.arch.pool.ntp.org).
Jul 24 19:23:07 dyb1b0k systemd-resolved[401]: Using degraded feature set UDP instead of UDP+EDNS0 for DNS server 198.18.0.2.
Jul 24 19:23:08 dyb1b0k systemd-resolved[401]: Clock change detected. Flushing caches.
Jul 24 19:23:08 dyb1b0k systemd-timesyncd[371]: Contacted time server 198.18.0.6:123 (3.arch.pool.ntp.org).
Jul 24 19:23:08 dyb1b0k systemd-timesyncd[371]: Initial clock synchronization to Thu 2025-07-24 19:23:08.237801 EEST.
Jul 24 19:23:08 dyb1b0k systemd[1]: systemd-hostnamed.service: Deactivated successfully.
Jul 24 19:23:38 dyb1b0k chrome[1697]: [1697:1716:0724/192338.401323:INFO:chrome/browser/extensions/extension_garbage_collector.cc:184] Garbage collection for extensions on file thread is complete.
Jul 24 19:23:58 dyb1b0k systemd-networkd[412]: Meta: Link DOWN
Jul 24 19:23:58 dyb1b0k systemd-networkd[412]: Meta: Lost carrier
Jul 24 19:24:17 dyb1b0k systemd-logind[681]: Power key pressed short.
Jul 24 19:24:17 dyb1b0k systemd-logind[681]: Powering off...
Jul 24 19:24:17 dyb1b0k systemd-logind[681]: System is powering down.

Your network/vpn kinda died and 20s later you soft-touched the power button.

ke46138 · 2025-07-28 09:25:25

I had the same problem, with the same graphics card, but with a different laptop (MSI Katana 17 GF76 11UCX-892XRU) and processor (Intel Core i5-11260H). Reinstalling arch linux solved this problem. I read the "Laptop issues" forum and saw that this problem exists on Nvidia RTX 2050 mobile and 3050 mobile (they use the same chip).

Sorry for my bad English, i used translator.

seth · 2025-07-28 12:39:48

"Reinstalling arch linux" solves exactly nothing.
You'll have changed something or it's just because of different package versions, but replicating the status quo ante can not lead to different results.

dybbuk · 2025-07-30 06:31:33

i think it's true that it's related to asusd/supergfxctl, i switched to TLP and it's gone

seth · 2025-07-30 07:47:45

Since asusctl is an AUR package, have you tried to update it (not re-install the local build but rebuild the package)?

dybbuk · 2025-07-30 13:36:53

sorry idk what you mean, all i do is `yay -Suy` lol

seth · 2025-07-30 13:46:08

https://wiki.archlinux.org/title/Makepkg
https://wiki.archlinux.org/title/Arch_User_Repository

Also "-Suy" is morally and ethically wrong, https://bbs.archlinux.org/viewtopic.php … 5#p2082235

dybbuk · 2025-07-31 22:07:08

nvm i just got the same issue even with tlp

http://0x0.st/87-Z.txt

i noticed that i get it when am doing some gpu-heavy stuff like running comfyui

seth · 2025-08-01 07:32:48

Does it remain stable if you explicitly DISable RTD3, https://wiki.archlinux.org/title/PRIME# … Management ?

dybbuk · 2025-08-03 11:22:14

i think disabling RTD3 fixed it but i need RTD3 becuase i'am on a hybrid laptop and i need the gpu to go suspended when not used to save bettery

seth · 2025-08-03 14:43:34

https://wiki.archlinux.org/title/NVIDIA … P_firmware ?

dybbuk · 2025-08-04 00:02:46

it fixed the issue thank u

i thought nvidia fixed that...

seth · 2025-08-04 06:40:26

Possibly in some development version, there's not been many versions w/ GSP enabled by default around and nvidia will likely focus on issues w/ newer GPUs mandating nvidia-open (where disabling the GSP is not an option)

Please always remember to mark resolved threads by editing your initial posts subject - so others will know that there's no task left, but maybe a solution to find.
Thanks.

Arch Linux

#1 2025-07-24 07:05:34

[Solved] nvidia drivers not detecting GPU after suspending it

#2 2025-07-24 11:32:49

Re: [Solved] nvidia drivers not detecting GPU after suspending it

#3 2025-07-24 19:39:00

Re: [Solved] nvidia drivers not detecting GPU after suspending it

#4 2025-07-24 19:50:50

Re: [Solved] nvidia drivers not detecting GPU after suspending it

#5 2025-07-24 19:52:07

Re: [Solved] nvidia drivers not detecting GPU after suspending it

#6 2025-07-24 19:55:10

Re: [Solved] nvidia drivers not detecting GPU after suspending it

#7 2025-07-27 20:45:01

Re: [Solved] nvidia drivers not detecting GPU after suspending it

#8 2025-07-27 22:05:55

Re: [Solved] nvidia drivers not detecting GPU after suspending it

#9 2025-07-28 00:17:49

Re: [Solved] nvidia drivers not detecting GPU after suspending it

#10 2025-07-28 06:54:37

Re: [Solved] nvidia drivers not detecting GPU after suspending it

#11 2025-07-28 09:25:25

Re: [Solved] nvidia drivers not detecting GPU after suspending it

#12 2025-07-28 12:39:48

Re: [Solved] nvidia drivers not detecting GPU after suspending it

#13 2025-07-30 06:31:33

Re: [Solved] nvidia drivers not detecting GPU after suspending it

#14 2025-07-30 07:47:45

Re: [Solved] nvidia drivers not detecting GPU after suspending it

#15 2025-07-30 13:36:53

Re: [Solved] nvidia drivers not detecting GPU after suspending it

#16 2025-07-30 13:46:08

Re: [Solved] nvidia drivers not detecting GPU after suspending it

#17 2025-07-31 22:07:08

Re: [Solved] nvidia drivers not detecting GPU after suspending it

#18 2025-08-01 07:32:48

Re: [Solved] nvidia drivers not detecting GPU after suspending it

#19 2025-08-03 11:22:14

Re: [Solved] nvidia drivers not detecting GPU after suspending it

#20 2025-08-03 14:43:34

Re: [Solved] nvidia drivers not detecting GPU after suspending it

#21 2025-08-04 00:02:46

Re: [Solved] nvidia drivers not detecting GPU after suspending it

#22 2025-08-04 06:40:26

Re: [Solved] nvidia drivers not detecting GPU after suspending it

Board footer