You are not logged in.
I'm officially confused now.... Through sampling I tested several versions of the LTS. Now i'm at the latest one: 6.18.57-2 and everything works!
You see what i'm dealing with???
Now, a note. Before starting to test various versions I removed linux-lts-headers installation so now i'm running without those. Not sure i they can make any difference.....
Offline
Headers are the problem somehow!
After the previous post I have installed linux-lts-headers.
Rebooted -> laptop got stuck
I rebooted again to level 3, removed linux-lts-headers
Rebooted again and here it works.
Not sure how or why headers could cause the problem!
Offline
The nvidia-dkms module (or any other dkms module!) depends on the headers (but only to build), so maybe you actually *had* that installed or there's some other out of tree module that causes this?
pacman -Qs dkmsOffline
pacman -Qs dkms
local/dkms 3.4.4-1
Dynamic Kernel Modules System
local/virtualbox-host-dkms 7.2.20-1
VirtualBox Host kernel modules sourcesOffline
So… it's the virtualbox modules??
Do they get loaded? What if you remove or blacklist them?
Offline
Hmmm you know what's funny?
1) I have created
sudo nano /etc/modprobe.d/blacklist-virtualbox.confwith this content:
blacklist vboxdrv
blacklist vboxnetadp
blacklist vboxnetflt
blacklist vboxpci2) Installed linux-lts-headers
3) Re-made lts image
4) Rebooted to 6.18 level 3. Confirmed using lsmod no vbox module is loaded
5) Tried running lightdm
6) Got stuck again.
Ok, then I
1) Removed blacklist file
2) Removed headers
3) Rebooted to level 3
4) Re-made image
5) Rebooted again to level 3
6) Tried starting lightdm: got stuck again hhahahahh
Voodoo ![]()
Now, one thing. I could be important. I can't remember exactly in what order, possibly after the final succesfull 6.18 boot, but I have installed acpi_call-lts (which removed acpi_call-dkms which I had installed before that). Not sure if that could have made a difference. I do need both acpi_call and vbox module to be honest.
Offline
Any chance the previous alignment w/ the headers has been a fluke?
Did you try w/o acpi_call (this actually makes more sense)?
Can you restore the working state at all?
Offline
First of all sorry guys for late reply, I had to handle some urgent private things these days.
Ok, Seth you are on to something!
I removed "acpi_call-lts". The hook regenerated 6.18 image. I booted - it worked.
Then I re-installed that package, rebooted 6.18 - it failed.
Then I removed it again, rebooted and 6.18 works.
Unfortunately I need "acpi_call" because I use this script daily to cut my NVidia card power (to keep the laptop silent and extend battery life for something like +20%):
#!/bin/bash
if lsmod | grep -q acpi_call; then
methods="
\_SB.PCI0.PEG0.PEGP._OFF
"
for m in $methods; do
echo -n "Trying $m: "
echo $m > /proc/acpi/call
result=$(cat /proc/acpi/call)
case "$result" in
Error*)
echo "failed"
;;
*)
echo "works!"
# break # try out outher methods too
;;
esac
done
else
echo "The acpi_call module is not loaded, try running 'modprobe acpi_call' or 'insmod acpi_call.ko' as root"
exit 1
fiWhat's the alternative?
Offline
I'd first and foremost disable that script to see whether it (cutting Nvidia power at this point and time) is the a tual cause.
Then in case explore whether calling it at some other Point (earlier or later) works reliably.
Offline
Ah, that script is run manually only after I login. That's not a problem.
As that script I'm using for at least 7 years now, and as that script is the only reason I had to have "acpi_call" I've decided to try different approach. At the moment i'm investigating this:
sudo tee /etc/udev/rules.d/80-nvidia-runtime-pm.rules >/dev/null <<'EOF'
ACTION=="add", SUBSYSTEM=="pci", KERNEL=="0000:01:00.0", TEST=="power/control", ATTR{power/control}="auto"
EOFand it seems to be working for now. So I won't need acpi_call and the script anymore. I'm just curious how come acpi_call itself could cause this laptop/kernel issue....
Offline
I don't really believe that merely loading the module will cause this.
Do you have optimus-manager, bbswitch or so installed
You can monitor access to the acpi file,
https://wiki.archlinux.org/title/Audit_framework
Offline
Ok, this is again something new for me! Thanks, I'll check it.
I don't have any of "optimus-manager, bbswitch". I don't switch between graphics cards really...
To be honest, with this udev rule seems as if thing settled for now.
Offline
this method only works for pre-turing cards so please ignore this as advice for gtx 1650 card.
you should really try bbswitch for disabling the nvidia card, ive been using it for years on one of my laptops for this purpose, and recently i noticed that laptop getting hotter than normal and this thread promtped me to check it, turns out bbswitch isnt in the main repos anymore and hasnt been working, i installed it from the aur and its now disabled again.
all you need to do is install bbswitch and use these config files:
/etc/modprobe.d/bbswitch.conf
options bbswitch load_state=0 unload_state=1/etc/modules-load.d/bbswitch.conf
bbswitchEDIT: forgot the modules-load conf, fixed now
EDIT: after further reading it seems that bbswitch wont work with turing and later cards and the RTD3 power management that your using is the way to go, only downside is you need to install the nvidia drivers even though your never going to use them, have updated my post to suit
Last edited by jonno2002 (2026-10-28 19:03:35)
Offline
@bope, you're currently doing https://wiki.archlinux.org/title/PRIME# … Management
If that's fine w/ you it's actually a common approach and allows you to dynamically use the nvidia GPU.
In that case and 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.
The only pending "question" would be what else would have invoked acpi_call unconditionally (and early) - but you don't /need/ an answer to that ![]()
Offline
Ok, I have marked my thread as solved because, you are right - it's functionally solved ![]()
So, I don't use Prime or anything else. When I got this laptop my idea was to have a dual boot Linux so that on Windows, once in a blue moon, I can play some game decently. You know, I'm one of those guys who typically play Caesar 3 on Linux (lol) or Age Of Empires lol. Once every 6-12 months I dual boot to Windows and spend 2 days straight playing something like World Of Tanks or something more serious. That's the reason I didn't want to know my GPU exists while I'm on Linux ![]()
When it comes to acpi_call I do believe Audit_framework could help me with that. I'll investigate it as I'm curious how come this started happening all of a sudden
Offline