You are not logged in.
I'm new to this so, sorry in advance for any missing information or poor communication.
So, this is the first time I've installed linux outside of a VM and as of now, the system does not turn on after any of the "idle events" and always end up requiring hard reboot. It wakes up if left idle for less than 2 minutes~ after screen turns off but after that it blacks out.
Here's my system information:
OS: Arch Linux x86_64
Host: 83JC (LOQ 15ARP9)
Kernel: Linux 6.17.13-hardened1-2-hardened
Uptime: 8 mins
Packages: 831 (pacman)
Shell: bash 5.3.9
Display (AUO7EAD): 1920x1080 in 16", 60 Hz [Built-in]
DE: KDE Plasma 6.5.5
WM: KWin (Wayland)
WM Theme: Breeze
Theme: Breeze (Light) [Qt], Breeze-Dark [GTK2], Breeze [GTK3]
Icons: breeze [Qt], breeze [GTK2/3/4]
Font: Noto Sans (10pt) [Qt], Noto Sans (10pt) [GTK2/3/4]
Cursor: KDE_Classic (24px)
Terminal: konsole 25.12.1
CPU: AMD Ryzen 5 7235HS (8) @ 4.25 GHz
GPU: NVIDIA GeForce RTX 4050 Max-Q / Mobile [Discrete]
Memory: 2.00 GiB / 15.30 GiB (13%)
Swap: 0 B / 7.65 GiB (0%)
current idle state is s2idle
I've tried individually enabling nvidia sleep/hibernate/resume services. Even tried switching TTY but doesn't work, since screen is off. I have read a few other posts about doing something with journalctl but i have not much idea about it to understand it. Any an every help is appreciated.
Offline
end up requiring hard reboot
Avoid that at all costs and see whether you can still switch the VT (ctrl+alt+F3), ssh into the system or reboot using the https://wiki.archlinux.org/title/Keyboa … el_(SysRq) + REISUB
Then see whether you're facing the issue when suspending from the multi-user.target (2nd link below) and w/ the normal or LTS kernel (instead of the hardened one)
You can also try https://aur.archlinux.org/packages?O=0&K=580xx and disabling the https://wiki.archlinux.org/title/NVIDIA … P_firmware
Offline
thanks for the sysrq tip.
No, I am unable to switch the VT, or maybe i did but can't know because the screen literally does not power on after sleep/suspend. The issue persists in multi-user.target.
Haven't tried other kernels YET (doing that next). Disabled GPG firmware but that didn't help either. Current driver version is 590.48.01.
Offline
just a hunch, sorry if i'm wrong
do you have?
nvidia.NVreg_PreserveVideoMemoryAllocations=1also I sus NVIDIA's PM is being aggressive in your case
echo "options nvidia NVreg_DynamicPowerManagement=0x02" | sudo tee /etc/modprobe.d/nvidia-pm.confrebuild initramfs
edit:
if not helped remove the changes esp rebuild initramfs, and keep the system at current state/stage.
Last edited by 5hridhyan (2026-01-26 09:39:35)
stay righteous too long in a broken world, and the crown you earn starts to rot
Online
Disabled GPG firmware but that didn't help either. Current driver version is 590.48.01.
nb. that you *cannot* disable GSP w/ nvidia-open, you'll *have* to use https://aur.archlinux.org/packages?O=0&K=580xx for that.
thanks for the sysrq tip
Can you use it to reboot out of that situation (+ REISUB, the S and the U are particularly important)
Because then you can post your complete system journal for the previous boot afterwards:
sudo journalctl -b -1 | curl -F 'file=@-' 0x0.stand it might hint at what's wrong.
Offline
just a hunch, sorry if i'm wrong
do you have?nvidia.NVreg_PreserveVideoMemoryAllocations=1
Is that an option in a modprobe config file? if yes, no i do not have it.
as for the other suspicion, i did what you asked and updated the initramfs. it did not help, unfortunately.
Offline
https://wiki.archlinux.org/title/Kernel_parameters but this is actually the default on current nvidia drivers.
DPM is mostly relevant for prime, https://wiki.archlinux.org/title/PRIME#NVIDIA - and depends on the actual GPU generation and you seem to have an Ada chip.
Offline
nb. that you *cannot* disable GSP w/ nvidia-open, you'll *have* to use https://aur.archlinux.org/packages?O=0&K=580xx for that.
By disable, i just meant i set the module parameter as NVreg_EnableGpuFirmware=0 in nvidia.conf. sorry for mislabeling.
Here's the text file journalctl produced:
http://0x0.st/PNsM.txt
Offline
Jan 26 20:28:51 snofaowfall kernel: NVRM: loading NVIDIA UNIX x86_64 Kernel Module 580.126.09 Wed Jan 7 22:59:56 UTC 2026Jan 26 20:43:07 snofaowfall systemd[1]: Reached target Sleep.
Jan 26 20:43:07 snofaowfall kscreenlocker_greet[11275]: QWaylandGLContext: Failed to create the decorations EGLContext. Decorations will not be drawn.
Jan 26 20:43:07 snofaowfall systemd[1]: Starting NVIDIA system suspend actions...
Jan 26 20:43:07 snofaowfall suspend[11298]: nvidia-suspend.service
Jan 26 20:43:07 snofaowfall logger[11298]: <13>Jan 26 20:43:07 suspend: nvidia-suspend.service
Jan 26 20:43:08 snofaowfall kernel: BUG: using __this_cpu_read() in preemptible [00000000] code: systemd-logind/490
Jan 26 20:43:08 snofaowfall kernel: caller is get_stack_info_noinstr+0x86/0x110
Jan 26 20:43:08 snofaowfall kernel: CPU: 2 UID: 0 PID: 490 Comm: systemd-logind Tainted: P OE 6.17.13-hardened1-2-hardened #1 PREEMPT(full) 08beba15f32e558254bd94307d45611cea050d9c
Jan 26 20:43:08 snofaowfall kernel: Tainted: [P]=PROPRIETARY_MODULE, [O]=OOT_MODULE, [E]=UNSIGNED_MODULE
Jan 26 20:43:08 snofaowfall kernel: Hardware name: LENOVO 83JC/LNVNB161216, BIOS PQCN26WW 09/19/2025
Jan 26 20:43:08 snofaowfall kernel: Call Trace:
Jan 26 20:43:08 snofaowfall kernel: <TASK>
Jan 26 20:43:08 snofaowfall kernel: dump_stack_lvl+0x5d/0x80
Jan 26 20:43:08 snofaowfall kernel: check_preemption_disabled+0xe5/0xf0
Jan 26 20:43:08 snofaowfall kernel: get_stack_info_noinstr+0x86/0x110
Jan 26 20:43:08 snofaowfall kernel: get_stack_info+0x2c/0x80
Jan 26 20:43:08 snofaowfall kernel: stack_access_ok+0x51/0xa0
Jan 26 20:43:08 snofaowfall kernel: unwind_next_frame+0x1cb/0x7b0
Jan 26 20:43:08 snofaowfall kernel: ? _nv003169kms+0x42/0x50 [nvidia_modeset 5ee65f74159052afeb6234c9669116f7176b4b25]
Jan 26 20:43:08 snofaowfall kernel: ? __pfx_stack_trace_consume_entry+0x10/0x10
Jan 26 20:43:08 snofaowfall kernel: arch_stack_walk+0xa6/0x110
Jan 26 20:43:08 snofaowfall kernel: ? _nv003169kms+0x42/0x50 [nvidia_modeset 5ee65f74159052afeb6234c9669116f7176b4b25]
Jan 26 20:43:08 snofaowfall kernel: stack_trace_save+0x4d/0x70
Jan 26 20:43:08 snofaowfall kernel: __kfence_alloc+0xb7/0x6f0
Jan 26 20:43:08 snofaowfall kernel: ? __kfence_alloc+0xb7/0x6f0
Jan 26 20:43:08 snofaowfall kernel: ? __kmalloc_noprof+0x52b/0x560
Jan 26 20:43:08 snofaowfall kernel: ? os_alloc_mem+0x108/0x120 [nvidia 6bcd1ae096542ab8c03360e3a33a5a5e006a1898]
Jan 26 20:43:08 snofaowfall kernel: ? _nv015303rm+0x34/0x50 [nvidia 6bcd1ae096542ab8c03360e3a33a5a5e006a1898]
Jan 26 20:43:08 snofaowfall kernel: ? _nv015305rm+0x2b/0xd0 [nvidia 6bcd1ae096542ab8c03360e3a33a5a5e006a1898]
Jan 26 20:43:08 snofaowfall kernel: ? _nv016363rm+0x1c/0x90 [nvidia 6bcd1ae096542ab8c03360e3a33a5a5e006a1898]
Jan 26 20:43:08 snofaowfall kernel: ? _nv059426rm+0x65/0xb0 [nvidia 6bcd1ae096542ab8c03360e3a33a5a5e006a1898]
Jan 26 20:43:08 snofaowfall kernel: ? _nv054160rm+0x20f/0x360 [nvidia 6bcd1ae096542ab8c03360e3a33a5a5e006a1898]
Jan 26 20:43:08 snofaowfall kernel: ? _nv054179rm+0x15f/0x1e0 [nvidia 6bcd1ae096542ab8c03360e3a33a5a5e006a1898]
Jan 26 20:43:08 snofaowfall kernel: ? _nv015927rm+0x345/0x680 [nvidia 6bcd1ae096542ab8c03360e3a33a5a5e006a1898]
Jan 26 20:43:08 snofaowfall kernel: ? _nv054134rm+0x69/0xd0 [nvidia 6bcd1ae096542ab8c03360e3a33a5a5e006a1898]
Jan 26 20:43:08 snofaowfall kernel: ? _nv014193rm+0x86/0xa0 [nvidia 6bcd1ae096542ab8c03360e3a33a5a5e006a1898]
Jan 26 20:43:08 snofaowfall kernel: ? _nv000652rm+0x5e/0x70 [nvidia 6bcd1ae096542ab8c03360e3a33a5a5e006a1898]
Jan 26 20:43:08 snofaowfall kernel: ? rm_kernel_rmapi_op+0x167/0x273 [nvidia 6bcd1ae096542ab8c03360e3a33a5a5e006a1898]
Jan 26 20:43:08 snofaowfall kernel: ? nvkms_call_rm+0x4c/0x80 [nvidia_modeset 5ee65f74159052afeb6234c9669116f7176b4b25]
Jan 26 20:43:08 snofaowfall kernel: ? _nv003169kms+0x42/0x50 [nvidia_modeset 5ee65f74159052afeb6234c9669116f7176b4b25]
Jan 26 20:43:08 snofaowfall kernel: __kmalloc_noprof+0x52b/0x560Test the behavior w/ the non-hardened kernel and if it fails there as well try to explicitly disable https://wiki.archlinux.org/title/NVIDIA … er_suspend (kernel parameter AND userspace services)
Offline
sorry for confusing earlier, and from the logs I recommend you test with linux or linux-lts or linux-zen kernels....
if that also did'nt work then I would suspect the suspend modes
edit: repeated post by mistake, should have refreshed the page, missed recent replies, sorry
Last edited by 5hridhyan (2026-01-26 15:57:48)
stay righteous too long in a broken world, and the crown you earn starts to rot
Online
so... about the different kernel... i installed linux-lts but i am not able to install the linux nvidia driver. i just did the standard
yay -S nvidia-580xx-dkms nvidia-580xx-utils lib32-nvidia-580xx-utils and initialized the initramfs but the driver simply doesn't appear on this when tested with nvidia-smi.
Am i doing something wrong...?
Last edited by silverfos_ (2026-01-26 15:58:57)
Offline
I refreshed the page I hope i didnt missed any recent replies this time
have you installed "linux-lts-headers"? if not install it reboot then reinstall
yay -S nvidia-580xx-dkms nvidia-580xx-utils lib32-nvidia-580xx-utils and then check status with "dkms status" before that
sudo dkms autoinstallthen again regenerate and reboot to LTS and check again.
edit:
DKMS does not need mkinitcpio every time, It builds on kernel install only if headers exist
Last edited by 5hridhyan (2026-01-26 16:06:25)
stay righteous too long in a broken world, and the crown you earn starts to rot
Online
thank you for the information on linux-lts-headers. it worked!
now getting back to testing suspend on lts
screen did not turn on again. here's the txt file generated by the same command:
http://0x0.st/PNzb.txt
edit because I forgot to test manually disabling nvidia services. Upon doing so, system does sleep (kinda?) it doesn't go into full sleep. it just goes to the login screen and system is wide awake. of course upon closing lid screen turns off but it immediately turns on. to answer "how" i know it doesn't turn off, the led on the power button is the answer. Formerly when on windows, sleep causes led to blink slowly. in this case, it is always on and fan is spinning in usual speed as it does when system is on.
Last edited by silverfos_ (2026-01-26 17:17:09)
Offline
Latest journal ends in
Jan 26 21:48:34 snofaowfall systemd-sleep[1908]: Performing sleep operation 'suspend'...
Jan 26 21:48:34 snofaowfall kernel: PM: suspend entry (s2idle)
Jan 26 21:48:34 snofaowfall kernel: Filesystems sync: 0.150 secondsDid you reboot w/ the power button?
Do you have a journal for the sleep attempt w/ the services disabled?
(It is *paramount* to also add "nvidia.NVreg_PreserveVideoMemoryAllocations=0" to the https://wiki.archlinux.org/title/Kernel_parameters )
Offline
Did you reboot w/ the power button?
No, i rebooted using the sysrq key.
Do you have a journal for the sleep attempt w/ the services disabled?
(It is *paramount* to also add "nvidia.NVreg_PreserveVideoMemoryAllocations=0" to the https://wiki.archlinux.org/title/Kernel_parameters )
adding the line to modprobe.d considers it a bad line so i added to /etc/sysctl.d/nvidia.conf
here's the journalctl after sleep and unlocking (because like i said earlier, system doesn't properly sleep, just goes to lock screen):
http://0x0.st/PNMK.txt
Offline
adding the line to modprobe.d considers it a bad line so i added to /etc/sysctl.d/nvidia.conf
Because it's a kernel parameter
Put it into the kernel commandline, not somewhere on disk.
There's no attempt to suspend the system in http://0x0.st/PNMK.txt - how *exactly* did you go about disabling what services in particular?
Disabling the nvidia services might run you into corrupt textures or complaints about the mismatch between userspace services and kernel parameter, but there should absolutely be an *attempt* to sleep the system logged.
The plan is to
- correctly add the kernel parameter (and early)
- suspend the system
- wake
- fail
- reboot w/ sysrq+REISUB (don't scroll over the keys, give each stage some seconds)
- post the journal of the previous boot (covering the botched wakeup)
Offline
Put it into the kernel commandline, not somewhere on disk.
Oh ok sry, now i put it in /boot/loader/entries/linux-lts.conf (name changed cause its a bunch of letters)
how *exactly* did you go about disabling what services in particular?
Disabling the nvidia services might run you into corrupt textures or complaints about the mismatch between userspace services and kernel parameter, but there should absolutely be an *attempt* to sleep the system logged.
Don't know why that's happening. I disabled suspend, hibernate and resume service using sysctl disable nvidia-xxx.service
The plan is to
- correctly add the kernel parameter (and early)
- suspend the system
- wake
- fail
- reboot w/ sysrq+REISUB (don't scroll over the keys, give each stage some seconds)
- post the journal of the previous boot (covering the botched wakeup)
I edited kernel parameters as mentioned above. disabled all 3 services. I'll update after remaining steps.
Offline
cat /proc/cmdline
systemctl list-unit-filesOffline
Uh.. unfortunately, I will not be able to continue this conversation for a while... My laptop fell down because of a wobbly table... And it's gone for repairs. Sorry...
Edit: System back. I had to reinstall arch since the efi partition got corrupted mid clone at repair shop (fortunately not much important data in arch yet)
Coming back, new install still has same issue.
Last edited by silverfos_ (2026-01-29 14:28:41)
Offline
Here's the same steps as before:
current system specs:
OS: Arch Linux x86_64
Host: 83JC (LOQ 15ARP9)
Kernel: Linux 6.12.67-1-lts
Uptime: 27 mins
Packages: 779 (pacman)
Shell: bash 5.3.9
Display (AUO7EAD): 1920x1080 in 16", 144 Hz [Built-in]
DE: KDE Plasma 6.5.5
WM: KWin (Wayland)
WM Theme: Breeze
Theme: Breeze (Light) [Qt], Breeze [GTK2/3]
Icons: breeze [Qt], breeze [GTK2/3/4]
Font: Noto Sans (10pt) [Qt], Noto Sans (10pt) [GTK2/3/4]
Cursor: breeze (24px)
Terminal: konsole 25.12.1
CPU: AMD Ryzen 5 7235HS (8) @ 4.25 GHz
GPU: NVIDIA GeForce RTX 4050 Max-Q / Mobile [Discrete]
Memory: 2.30 GiB / 15.31 GiB (15%)
Swap: 0 B / 7.65 GiB (0%)
Disk (/): 10.65 GiB / 99.39 GiB (11%) - btrfsfirst activate sysrq key.
i have not disabled suspend/hibernate/resume services.
journalctl file:
https://paste.c-net.org/DavidCabbie
(0x0.st not working in my pc. same with few others. fortunately paste-c works)
Now i disable suspend/hibernate/resume using systemctl stop and disable commands.
editing kernel parameters in /boot/loader/entries/linux-lts.conf
rebooted for effects to take place.
cat /proc/cmdline
initrd=\initramfs-linux-lts.img root=PARTUUID=1e824ea3-5c55-4a6e-8b65-b08a5d707ae1 zswap.enabled=0 rootflags=subvol=@root rw rootfstype=btrfs nvidia.NVreg_PreserveVideoMemoryAllocations=0systemctl list-unit-fileshttps://paste.c-net.org/ArsonistStefano
testing suspend.
hard rebooted cause i forgot sysrq after reboot. retrying:
did correctly. here's the journlctl:
https://paste.c-net.org/FolkloreArterial
hopefully i didn't miss anything this time. hope to hear from you soon
Offline
https://paste.c-net.org/DavidCabbie ends in
Jan 29 20:01:40 snowfall systemd-sleep[8467]: User sessions remain unfrozen on explicit request ($SYSTEMD_SLEEP_FREEZE_USER_SESSIONS=0).
Jan 29 20:01:40 snowfall systemd-sleep[8467]: This is not recommended, and might result in unexpected behavior, particularly
Jan 29 20:01:40 snowfall systemd-sleep[8467]: in suspend-then-hibernate operations or setups with encrypted home directories.
Jan 29 20:01:40 snowfall systemd-sleep[8467]: Performing sleep operation 'suspend'...
Jan 29 20:01:40 snowfall kernel: PM: suspend entry (s2idle)https://paste.c-net.org/FolkloreArterial in
Jan 29 20:56:12 snowfall systemd[1]: Reached target Sleep.
Jan 29 20:56:12 snowfall systemd[1]: Starting System Suspend...
Jan 29 20:56:12 snowfall systemd-sleep[1864]: User sessions remain unfrozen on explicit request ($SYSTEMD_SLEEP_FREEZE_USER_SESSIONS=0).
Jan 29 20:56:12 snowfall systemd-sleep[1864]: This is not recommended, and might result in unexpected behavior, particularly
Jan 29 20:56:12 snowfall systemd-sleep[1864]: in suspend-then-hibernate operations or setups with encrypted home directories.
Jan 29 20:56:12 snowfall systemd-sleep[1864]: Performing sleep operation 'suspend'...
Jan 29 20:56:12 snowfall kernel: PM: suspend entry (s2idle)
Jan 29 20:56:12 snowfall kernel: Filesystems sync: 0.017 seconds
Jan 29 20:56:12 snowfall kscreenlocker_greet[1848]: QWaylandGLContext: Failed to create the decorations EGLContext. Decorations will not be drawn.if you rebooted w/ sysrq+REISUB you probably lost the root partition…
https://wiki.archlinux.org/title/Solid_ … leshooting - disable APST and use the software IOMMU.
Also
Jan 29 20:51:51 snowfall kernel: nvme0n1: p1 p2 p3 p4 p5 p6 assuming there's a parallel windows installation see the 3rd link below. Mandatory.
Disable it (it's NOT the BIOS setting!) and reboot windows and linux twice for voodo reasons.
Offline
I just felt like i didn't give REISUB enough time. so here's the file after redo. confirm if i should proceed with same steps as mentioned before:
https://paste.c-net.org/HattieShadowy
additionally, could you clarify if i need to turn on suspend services and remove the video memory allocation kernel parameter?
Last edited by silverfos_ (2026-01-29 16:58:52)
Offline
Jan 29 22:21:01 snowfall systemd[1]: Reached target Sleep.
Jan 29 22:21:01 snowfall systemd[1]: Starting System Suspend...
Jan 29 22:21:01 snowfall systemd-sleep[1842]: User sessions remain unfrozen on explicit request ($SYSTEMD_SLEEP_FREEZE_USER_SESSIONS=0).
Jan 29 22:21:01 snowfall systemd-sleep[1842]: This is not recommended, and might result in unexpected behavior, particularly
Jan 29 22:21:01 snowfall systemd-sleep[1842]: in suspend-then-hibernate operations or setups with encrypted home directories.
Jan 29 22:21:01 snowfall systemd-sleep[1842]: Performing sleep operation 'suspend'...
Jan 29 22:21:01 snowfall kernel: PM: suspend entry (s2idle)
Jan 29 22:21:01 snowfall kscreenlocker_greet[1821]: QWaylandGLContext: Failed to create the decorations EGLContext. Decorations will not be drawn.
Jan 29 22:21:01 snowfall kernel: Filesystems sync: 0.031 seconds
Jan 29 22:21:01 snowfall kscreenlocker_greet[1821]: QWaylandGLContext: Failed to create the decorations EGLContext. Decorations will not be drawn.
Jan 29 22:21:01 snowfall kscreenlocker_greet[1821]: QWaylandGLContext: Failed to create the decorations EGLContext. Decorations will not be drawn.If that's the result of a sysrq reboot you're loosing the root partition => see the nvme parameters and don't ignore the windows fast start condition (this isn't optional)
You can revert the VRAM preservation settings if you want, r/n it doesn't seem to be related (change had no impact and we're worried about the disk)
Offline
removed the vram option and added the ones mentioned in the link:
vme_core.default_ps_max_latency_us=0
iommu=softSleep is still dysfunctional. additionally, when booting i noticed this:
[2.904059] ACPI BIOS Error (bug): Could not resolve symbol [\_SB.PCI0.GPP0.PEGP.GPS.NVD1], AE_NOT_FOUND (20240827/psargs-332)
[2.904059] ACPI Error: Aborting method \_SB.PCI0.GPP0.PEGP.GPS due to previous error (AE_NOT_FOUND) (20240827/pspargs-529)
[2.904059] ACPI Error: Aborting method \_SB.PCI0.GPP0.PEGP._DSM due to previous error (AE_NOT_FOUND) (20240827/pspargs-529)
https://ibb.co/Xk8fspwm
as seen in the above image
Do you think I should elevate this to a kernel and hardware question in forums?
Last edited by silverfos_ (2026-01-30 12:20:23)
Offline
"nvme_core.default_ps_max_latency_us=0 iommu=soft"
What's the output of
cat /proc/cmdline?
Did the ACPI errors only appear after switching to the 580xx drivers (and are you currently still using those)?
Also, you're stubbornly ignoring the comments about
don't ignore the windows fast start condition (this isn't optional)
If windows is hibernating there's really no mystery about what's going on here.
As a reminder see the 3rd link below. Mandatory.
Disable it (it's NOT the BIOS setting!) and reboot windows and linux twice for voodo reasons.
Offline