Arch Linux

justdanyul

Member

Registered: 2011-09-29

Posts: 130

Kernel 5.18 and nvidia 515.43.04-6

I know these should work together, but, I just updated my system and X doesn't load. Haven't had the chance to look into it yet, as I got some work I need to do so, alas, I have to turn to my trusty Macbook.  Anyways, I just wanted to check if others are having issues, before troubleshooting later tonight.

Last edited by justdanyul (2022-05-29 17:06:37)

Morta

Member

Registered: 2019-07-07

Posts: 655

Re: Kernel 5.18 and nvidia 515.43.04-6

Have you a RAID and/or crypted device? So it's a major bug look to others threads. Downgrade your Kernel to 5.17.9 (linux & linux-headers and nvidia&nivida-utils) with pacman -U. The packages are under /var/lib/cache/pkg. Regenerate initramfs with mkinticpio -P and reboot. Wait till they fix this issues!

NYI

Member

Registered: 2022-05-28

Posts: 2

Re: Kernel 5.18 and nvidia 515.43.04-6

Yes, same happened to me and has been reported by others:

https://github.com/NVIDIA/open-gpu-kern … issues/256

Adding ibt=off to my kernel parameters allowed me to boot.

justdanyul

Member

Registered: 2011-09-29

Posts: 130

Re: Kernel 5.18 and nvidia 515.43.04-6

Have you a RAID and/or crypted device? So it's a major bug look to others threads. Downgrade your Kernel to 5.17.9 (linux & linux-headers and nvidia&nivida-utils) with pacman -U. The packages are under /var/lib/cache/pkg. Regenerate initramfs with mkinticpio -P and reboot. Wait till they fix this issues!

Nope, no raid and/or crypted dev on this box (but thanks for the heads up, I wont update my home server).

justdanyul

Member

Registered: 2011-09-29

Posts: 130

Re: Kernel 5.18 and nvidia 515.43.04-6

Yes, same happened to me and has been reported by others:

https://github.com/NVIDIA/open-gpu-kern … issues/256

Adding ibt=off to my kernel parameters allowed me to boot.

thanks, I'll give this a go later

logothete

Member

Registered: 2022-01-06

Posts: 44

Re: Kernel 5.18 and nvidia 515.43.04-6

Removing the nvidia-dkms driver and switching to the standard one at least lets me use arch, but not the NVIDIA gpu.  Setting ibt=off does NOTHING, for me. I note here that I am indeed using an encrypted device.

The gpu is still detected:

lspci | grep "NVIDIA"                       
0000:01:00.0 VGA compatible controller: NVIDIA Corporation GA106M [GeForce RTX 3060 Mobile / Max-Q] (rev a1)
0000:01:00.1 Audio device: NVIDIA Corporation Device 228e (rev a1)

Manually attempting to start the nvidia daemon fails:

systemctl status nvidia-persistenced.service


nvidia-persistenced.service - NVIDIA Persistence Daemon
     Loaded: loaded (/usr/lib/systemd/system/nvidia-persistenced.service; enabled; vendor preset: disabled)
     Active: failed (Result: exit-code) since Sun 2022-05-29 19:25:59 EDT; 3min 46s ago
    Process: 697 ExecStart=/usr/bin/nvidia-persistenced --user nvidia-persistenced (code=exited, status=1/FAILURE)
    Process: 770 ExecStopPost=/bin/rm -rf /var/run/nvidia-persistenced (code=exited, status=0/SUCCESS)
        CPU: 9ms

May 29 19:25:59 psyche systemd[1]: Starting NVIDIA Persistence Daemon...
May 29 19:25:59 psyche nvidia-persistenced[699]: Started (699)
May 29 19:25:59 psyche nvidia-persistenced[699]: Failed to query NVIDIA devices. Please ensure that the NVIDIA device files (/dev/nvidia*) exist, and that user 143 has read and write p>
May 29 19:25:59 psyche nvidia-persistenced[697]: nvidia-persistenced failed to initialize. Check syslog for more details.
May 29 19:25:59 psyche nvidia-persistenced[699]: Shutdown (699)
May 29 19:25:59 psyche systemd[1]: nvidia-persistenced.service: Control process exited, code=exited, status=1/FAILURE
May 29 19:25:59 psyche systemd[1]: nvidia-persistenced.service: Failed with result 'exit-code'.
May 29 19:25:59 psyche systemd[1]: Failed to start NVIDIA Persistence Daemon.

So, I go to check that device files exist and... no, they don't. Which leaves me unsure what to do about it.

I cleared my caches entirely in between kernel 5.18 and 5.17.9, so my sad self has to figure out how to revert without the cache. Wish me luck.

Last edited by logothete (2022-05-29 23:51:21)

seth

Member

Registered: 2012-09-03

Posts: 51,332

Re: Kernel 5.18 and nvidia 515.43.04-6

I cleared my caches entirely in between kernel 5.18 and 5.17.9, so my sad self has to figure out how to revert without the cache.

You can just install the LTS kernel in addition and in doubt do so from the install iso (boot it, arch-chroot into the installed system, don't forget to mount /boot and install linux-lts and nvidia-lts)

The please post a complete journal of a troublesome boot w/ "ibt=off" set.

Edit: … eg. "sudo journalctl -b -1" for the previous one …

Last edited by seth (2022-05-30 04:57:11)

---

How to upload text · How to boot w/o GUI · Disable Windows Fast-Start! · Fix your xinitrc

justdanyul

Member

Registered: 2011-09-29

Posts: 130

Re: Kernel 5.18 and nvidia 515.43.04-6

ibt=off "solved" the problem for me

an_hero

Member

Registered: 2022-06-06

Posts: 1

Re: Kernel 5.18 and nvidia 515.43.04-6

I rolled back the kernel and headers with `pacman -U linux-5.17.9.arch1-1-x86_64.pkg.tar.zst linux-headers-5.17.9.arch1-1-x86_64.pkg.tar.zst` in order to get my graphical UI back up. My question on behalf of me and all those affected, how do we know when it's safe to upgrade the kernel again?

seth

Member

Registered: 2012-09-03

Posts: 51,332

Re: Kernel 5.18 and nvidia 515.43.04-6

It is safe to update the kernel now, disable ibt and wait until the nvida driver catches up before you enable it.
Ftr:

I rolled back the kernel and headers with `pacman -U linux-5.17.9.arch1-1-x86_64.pkg.tar.zst linux-headers-5.17.9.arch1-1-x86_64.pkg.tar.zst`

will have broken your system because you now have an incompatible nvidia build. I assume you downgraded the nvidia package as well?

---

How to upload text · How to boot w/o GUI · Disable Windows Fast-Start! · Fix your xinitrc

loqs

Member

Registered: 2014-03-06

Posts: 17,386

Re: Kernel 5.18 and nvidia 515.43.04-6

nvidia-open 515.48.07-2 currently in testing has been patched.  nvidia-open requires Turing or newer.

ecruz1986

Member

From: Vitoria - ES, Brazil

Registered: 2021-08-30

Posts: 2

Re: Kernel 5.18 and nvidia 515.43.04-6

My nvidia graphics acceleration suddenly stopped working. Now I have screen tearing and cannot run any games. Xorg is using modesetting driver, not nvidia.

I don't know which package update was the culprit... Everything was working fine a week ago.

I have RTX2070. I have nvidia-dkms. I have tried nvidia-open-dkms but the symptoms are the same. Adding ibt=off to my kernel parameters changes nothing, same symptoms.

I have tried booting linux-lts, but even there the bug exists.

I'm not sure if this is related to the this thread or if it is another bug altogether...

Here we see that x11 is using modesetting, and has failed loading nvidia:

eduardo@ECRUZ-PC:~$ inxi -Gx
Graphics:
  Device-1: NVIDIA TU106 [GeForce RTX 2070] vendor: eVga.com. driver: nvidia
    v: 515.43.04 arch: Turing bus-ID: 01:00.0
  Device-2: Razer USA Kiyo Pro type: USB driver: snd-usb-audio,uvcvideo
    bus-ID: 2-4:2
  Display: x11 server: X.Org v: 21.1.3 with: Xwayland v: 22.1.2 driver: X:
    loaded: modesetting unloaded: vesa failed: nvidia
    gpu: nvidia,nvidia-nvswitch resolution: 2560x1440
  OpenGL: renderer: llvmpipe (LLVM 13.0.1 256 bits) v: 4.5 Mesa 22.1.1
    direct render: Yes

My Xorg.0.log:

[    10.574] (EE) NVIDIA: Failed to initialize the NVIDIA kernel module. Please see the
[    10.574] (EE) NVIDIA:     system's kernel log for additional error messages and
[    10.574] (EE) NVIDIA:     consult the NVIDIA README for details.

Under the kernel log I just see these lines as something that seems abnormal:

jun 06 22:29:09 ECRUZ-PC systemd-udevd[367]: nvidia: Process '/usr/bin/bash -c '/usr/bin/mknod -Z -m 666 /dev/nvidiactl c $(grep nvidia-frontend /proc/devices | cut -d \  -f 1) 255'' failed with exit code 1.
jun 06 22:29:09 ECRUZ-PC systemd-udevd[395]: nvidia_uvm: Process '/usr/bin/bash -c '/usr/bin/mknod -Z -m 666 /dev/nvidia-uvm c $(grep nvidia-uvm /proc/devices | cut -d \  -f 1) 0'' failed with exit code 1.
jun 06 22:29:09 ECRUZ-PC systemd-udevd[395]: nvidia_uvm: Process '/usr/bin/bash -c '/usr/bin/mknod -Z -m 666 /dev/nvidia-uvm-tools c $(grep nvidia-uvm /proc/devices | cut -d \  -f 1) 1'' failed with exit code 1.
jun 06 22:29:09 ECRUZ-PC systemd-udevd[367]: nvidia: Process '/usr/bin/bash -c 'for i in $(cat /proc/driver/nvidia/gpus/*/information | grep Minor | cut -d \  -f 4); do /usr/bin/mknod -Z -m 666 /dev/nvidia${i} c $(grep nvidia-frontend /proc/devices | cut -d \  -f 1) ${i}; done'' failed with exit code 1.

These nodes seem to be there:

eduardo@ECRUZ-PC:~$ ls /dev/nvi*
/dev/nvidia0  /dev/nvidiactl  /dev/nvidia-modeset  /dev/nvidia-uvm  /dev/nvidia-uvm-tools

There are nvidia kernel modules loaded:

eduardo@ECRUZ-PC:~$ lsmod | grep nvidia
i2c_nvidia_gpu         16384  0
nvidia_uvm           2732032  0
nvidia_drm             73728  1
nvidia_modeset       1433600  1 nvidia_drm
nvidia              44212224  2 nvidia_uvm,nvidia_modeset

loqs

Member

Registered: 2014-03-06

Posts: 17,386

Re: Kernel 5.18 and nvidia 515.43.04-6

ecruz1986 please start a new thread for your issue which is different to that of justdanyul's as you noted the solution for the original issue did not work for you.
Edit:
Please also include the full output of dmesg in the new thread.

Last edited by loqs (2022-06-07 02:30:10)