You are not logged in.
Hello! I recently switched to Arch from Ubuntu and I am loving it so far. Getting better FPS on native Windows games is great! One caveat is that I have an Nvidia 1080 Ti in my machine. Today when installing the latest zen-kernel, I ran into some issues.
- When installing the latest kernels (I still have the regular linux-kernel), my boot partition got full, and mkinitcpio (?) failed to generate the images.
- Now when I boot I see this in the boot logs:
Jun 06 17:12:39 arch-base kernel: [drm:nv_drm_atomic_commit [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Flip event timeout on head 0
Jun 06 17:12:41 arch-base kernel: [drm:nv_drm_event_callback [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to add encoder for NvKmsKapiDisplay 0x00000001
^ The above is followed by an Nvidia process dump with a lot of hex code.
I pasted the rest of the boot sequence here:
http://0x0.st/8g1Z.txt
And my secondary display doesn't work in Wayland anymore, however my third display, which is connected via Displayport chaining to the second one works.
It doesn't show up in KDE Display config,
SDDM correctly outputs to all 3 displays, X11 as well. It is just when I login to Wayland that it doesnt show any image. The system even beeps when I turn it off/on.
I removed the normal kernel, and reinstalled the nvidia-dkms + nvidia-utils packages.
I have tried to regenerate the images in /boot/ with mkinitcpio -P but the issue persists.
I even tried downgrading to the last kernel that it worked on (6.14.9) with a mkinitcpio retrigger. Same thing with the normal kernel.
The solution seems to be downgrading to the last Nvidia driver (570.86.16-2). I'm not sure how to report this as a Wayland bug (if it even is one), or perhaps a bug with the new Nvidia driver, so I thought I'd begin by posting it here.
Has anyone else encountered this issue and maybe even solved it? Staying on the last driver for now!
Last edited by phpTANKER (2025-06-06 16:24:44)
Offline
When installing the latest kernels (I still have the regular linux-kernel), my boot partition got full, and mkinitcpio (?) failed to generate the images.
Is this the status quo?
You could remove the kms hook and not explicitly add the nvidia modules to the initramfs since the GSP firmware will considerably blow up the initramfs.
Online
When installing the latest kernels (I still have the regular linux-kernel), my boot partition got full, and mkinitcpio (?) failed to generate the images.
Is this the status quo?
You could remove the kms hook and not explicitly add the nvidia modules to the initramfs since the GSP firmware will considerably blow up the initramfs.
The status quo seems to be to stay on 570 drivers for me, regardless of if generating the images with dkms works or not, the failing image is the fallback.
I tried removing the regular linux kernel and upgrade the drivers to make sure that was not the issue.
Offline
No, I meant whether the status quo is that your /boot partition is underdimensioned and you're running out of space…
the failing image is the fallback
https://wiki.archlinux.org/title/Mkinit … generation
If that's not it, please post your complete system journal for an affected boot, eg.
sudo journalctl -b -1 | curl -F 'file=@-' 0x0.st
for the previous ("-1") one.
Online
No, I meant whether the status quo is that your /boot partition is underdimensioned and you're running out of space…
the failing image is the fallback
https://wiki.archlinux.org/title/Mkinit … generation
If that's not it, please post your complete system journal for an affected boot, eg.
sudo journalctl -b -1 | curl -F 'file=@-' 0x0.st
for the previous ("-1") one.
I think I introduced this issue by keeping linux / linux-headers after installing linux-zen, I have since then confirmed that it does in fact not cause the Nvidia driver issue presented in 575. The affected boot with the error message is in my original post, here is a clean boot with the 570 driver: http://0x0.st/8glT.txt
Last edited by phpTANKER (2025-06-08 17:50:50)
Offline
linux-zen is for now still at 6.14.10.zen1-1, that's not a problem here.
For clarification, after updating to 575xx your /boot partition is *not* full? And
and mkinitcpio (?) failed to generate the images
that also no longer happens?
Please post the pacman log covering the 575xx installation.
Online
linux-zen is for now still at 6.14.10.zen1-1, that's not a problem here.
For clarification, after updating to 575xx your /boot partition is *not* full? Andand mkinitcpio (?) failed to generate the images
that also no longer happens?
Please post the pacman log covering the 575xx installation.
Exactly, the images build without issues when there is space in boot, here is pacman upgrading Nvidia: http://0x0.st/8gG1.txt
And here is a boot directly after successfully installing the drivers (but the 2nd display not working): https://0x0.st/8gG2.txt
Offline
Jun 09 01:35:42 arch-base kernel: NVRM: loading NVIDIA UNIX x86_64 Kernel Module 575.57.08 Sat May 24 07:21:16 UTC 2025
Jun 09 01:35:42 arch-base kernel: nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms 575.57.08 Sat May 24 06:52:56 UTC 2025
Jun 09 01:35:42 arch-base kernel: nvidia_uvm: module uses symbols nvUvmInterfaceDisableAccessCntr from proprietary module nvidia, inheriting taint.
Jun 09 01:35:42 arch-base kernel: [drm] [nvidia-drm] [GPU ID 0x00000100] Loading driver
Jun 09 01:35:42 arch-base kernel: [drm:nv_drm_load [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to add connector for NvKmsKapiDisplay 0x00000800
Jun 09 01:35:42 arch-base kernel: [drm:nv_drm_load [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to get dynamic displays
Jun 09 01:35:42 arch-base kernel: [drm] Initialized nvidia-drm 0.0.0 for 0000:01:00.0 on minor 1
Jun 09 01:35:42 arch-base kernel: nvidia 0000:01:00.0: vgaarb: deactivate vga console
Jun 09 01:35:42 arch-base kernel: fbcon: nvidia-drmdrmfb (fb0) is primary device
Jun 09 01:35:42 arch-base kernel: [drm:nv_drm_atomic_commit [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Flip event timeout on head 0
Jun 09 01:35:42 arch-base kernel: nvidia 0000:01:00.0: [drm] fb0: nvidia-drmdrmfb frame buffer device
Just a hunch, add "nvidia-drm.modeset=1 nvidia-drm.fbdev=0" to the https://wiki.archlinux.org/title/Kernel_parameters
The former will block
Jun 09 01:35:42 arch-base kernel: simple-framebuffer simple-framebuffer.0: [drm] fb0: simpledrmdrmfb frame buffer device
(which you're most likely currently using, so make sure you know how to edit the kernel commandline parameters AT THE BOOTLOADER) the latter defaults to "1" but is reported to not work w/ older GPUs.
Online
x-ref, but that one's nvidia-open: https://bbs.archlinux.org/viewtopic.php … 1#p2245471
Online
Jun 09 01:35:42 arch-base kernel: NVRM: loading NVIDIA UNIX x86_64 Kernel Module 575.57.08 Sat May 24 07:21:16 UTC 2025 Jun 09 01:35:42 arch-base kernel: nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms 575.57.08 Sat May 24 06:52:56 UTC 2025 Jun 09 01:35:42 arch-base kernel: nvidia_uvm: module uses symbols nvUvmInterfaceDisableAccessCntr from proprietary module nvidia, inheriting taint. Jun 09 01:35:42 arch-base kernel: [drm] [nvidia-drm] [GPU ID 0x00000100] Loading driver Jun 09 01:35:42 arch-base kernel: [drm:nv_drm_load [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to add connector for NvKmsKapiDisplay 0x00000800 Jun 09 01:35:42 arch-base kernel: [drm:nv_drm_load [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to get dynamic displays Jun 09 01:35:42 arch-base kernel: [drm] Initialized nvidia-drm 0.0.0 for 0000:01:00.0 on minor 1 Jun 09 01:35:42 arch-base kernel: nvidia 0000:01:00.0: vgaarb: deactivate vga console Jun 09 01:35:42 arch-base kernel: fbcon: nvidia-drmdrmfb (fb0) is primary device Jun 09 01:35:42 arch-base kernel: [drm:nv_drm_atomic_commit [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Flip event timeout on head 0 Jun 09 01:35:42 arch-base kernel: nvidia 0000:01:00.0: [drm] fb0: nvidia-drmdrmfb frame buffer device
Just a hunch, add "nvidia-drm.modeset=1 nvidia-drm.fbdev=0" to the https://wiki.archlinux.org/title/Kernel_parameters
The former will blockJun 09 01:35:42 arch-base kernel: simple-framebuffer simple-framebuffer.0: [drm] fb0: simpledrmdrmfb frame buffer device
(which you're most likely currently using, so make sure you know how to edit the kernel commandline parameters AT THE BOOTLOADER) the latter defaults to "1" but is reported to not work w/ older GPUs.
I will give this a go tonight, thanks for the tips so far!
Offline
Jun 09 01:35:42 arch-base kernel: NVRM: loading NVIDIA UNIX x86_64 Kernel Module 575.57.08 Sat May 24 07:21:16 UTC 2025 Jun 09 01:35:42 arch-base kernel: nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms 575.57.08 Sat May 24 06:52:56 UTC 2025 Jun 09 01:35:42 arch-base kernel: nvidia_uvm: module uses symbols nvUvmInterfaceDisableAccessCntr from proprietary module nvidia, inheriting taint. Jun 09 01:35:42 arch-base kernel: [drm] [nvidia-drm] [GPU ID 0x00000100] Loading driver Jun 09 01:35:42 arch-base kernel: [drm:nv_drm_load [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to add connector for NvKmsKapiDisplay 0x00000800 Jun 09 01:35:42 arch-base kernel: [drm:nv_drm_load [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to get dynamic displays Jun 09 01:35:42 arch-base kernel: [drm] Initialized nvidia-drm 0.0.0 for 0000:01:00.0 on minor 1 Jun 09 01:35:42 arch-base kernel: nvidia 0000:01:00.0: vgaarb: deactivate vga console Jun 09 01:35:42 arch-base kernel: fbcon: nvidia-drmdrmfb (fb0) is primary device Jun 09 01:35:42 arch-base kernel: [drm:nv_drm_atomic_commit [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Flip event timeout on head 0 Jun 09 01:35:42 arch-base kernel: nvidia 0000:01:00.0: [drm] fb0: nvidia-drmdrmfb frame buffer device
Just a hunch, add "nvidia-drm.modeset=1 nvidia-drm.fbdev=0" to the https://wiki.archlinux.org/title/Kernel_parameters
The former will blockJun 09 01:35:42 arch-base kernel: simple-framebuffer simple-framebuffer.0: [drm] fb0: simpledrmdrmfb frame buffer device
(which you're most likely currently using, so make sure you know how to edit the kernel commandline parameters AT THE BOOTLOADER) the latter defaults to "1" but is reported to not work w/ older GPUs.
I just tried adding the parameters to systemd-boot (and later refind to see if it would make a difference). I seem to get the same errors on boot:
http://0x0.st/8Eiq.txt
Offline
Command line: initrd=\initramfs-linux-zen.img root=PARTUUID=a04c7f64-deb7-4c0a-b7f5-f955c8666d07 zswap.enabled=0 rw rootfstype=ext4
You did not get them added. Which boot manager are you actually using?
Online
Command line: initrd=\initramfs-linux-zen.img root=PARTUUID=a04c7f64-deb7-4c0a-b7f5-f955c8666d07 zswap.enabled=0 rw rootfstype=ext4
You did not get them added. Which boot manager are you actually using?
I noticed as well, strange, using reFind to boot arch with systemd-boot, since I have a windows install on another disk. Pressed e and added the parameters without quotes on systemd, in reFind when I press tab twice I got an empty text screen where I just entered the 2 params and pressed enter.
Offline
using reFind to boot arch
https://wiki.archlinux.org/title/REFInd … parameters
with systemd-boot
errr… what? You're chaining those bootloaders?
https://wiki.archlinux.org/title/System … ng_loaders
ls -lR /[your esp here]
since I have a windows install on another disk
3rd link below. Mandatory.
Disable it (it's NOT the BIOS setting!) and reboot windows and linux twice for voodo reasons.
While at it and because of the GSP crash in the other thread, add "nvidia.NVreg_EnableGpuFirmware=0" to the list.
Online
Same thing happened here today. dkms status says the nvidia drivers are installed yet they do not work. I think it has to do something with the kernel update.
Offline
dkms status says the nvidia drivers are installed
Please don't paraphrase, https://bbs.archlinux.org/viewtopic.php?id=57855
lspci -k
You might be running into nova_core, now, https://bbs.archlinux.org/viewtopic.php … 9#p2245579
Online
using reFind to boot arch
https://wiki.archlinux.org/title/REFInd … parameters
with systemd-boot
errr… what? You're chaining those bootloaders?
https://wiki.archlinux.org/title/System … ng_loadersls -lR /[your esp here]
since I have a windows install on another disk
3rd link below. Mandatory.
Disable it (it's NOT the BIOS setting!) and reboot windows and linux twice for voodo reasons.While at it and because of the GSP crash in the other thread, add "nvidia.NVreg_EnableGpuFirmware=0" to the list.
I should probably remove systemd-boot but I'm a bit afraid I might break my boot process, my bootctl info looks a bit weird:
http://0x0.st/8EQz.txt
However I will attempt to get the kernel parameters into my boot, I'd just rather do it temporarily than overriding any boot entries and potentially breaking the system.
EDIT: I created a manual boot entry in refind that successfully boots Arch now, so bootctl looks better! On to testing out the newer drivers with the 3 parameters.
Last edited by phpTANKER (2025-06-11 15:00:55)
Offline
using reFind to boot arch
https://wiki.archlinux.org/title/REFInd … parameters
with systemd-boot
errr… what? You're chaining those bootloaders?
https://wiki.archlinux.org/title/System … ng_loadersls -lR /[your esp here]
since I have a windows install on another disk
3rd link below. Mandatory.
Disable it (it's NOT the BIOS setting!) and reboot windows and linux twice for voodo reasons.While at it and because of the GSP crash in the other thread, add "nvidia.NVreg_EnableGpuFirmware=0" to the list.
Here is a non-chained reFind boot with the latest drivers and all 3 parameters added, same issue:
https://0x0.st/8EQv.txt
Offline
Curveball: what if you remove the nvidia modules (and/or kms hook) from the initramfs?
Online
Curveball: what if you remove the nvidia modules (and/or kms hook) from the initramfs?
If I remove the nvidia modules from mkinitcpio.conf, is it easy to have them loaded/built outside? I'm not too familiar with this process.
I found some threads on Reddit where people are having similar issues with the latest driver:
https://www.reddit.com/r/archlinux/comm … ernel_615/
https://www.reddit.com/r/archlinux/comm … l_version/
I tried blacklisting nova_core but the issue still persists.
Offline
nova_core did prevent the nvidia driver from being loaded at all.
If you remove the modules from the initramfs, they'll simply be loaded slightly later.
There was a thread w/ similar error, but an eGPU that wasn't fully ready when the driver wanted to talk to it.
Online
nova_core did prevent the nvidia driver from being loaded at all.
If you remove the modules from the initramfs, they'll simply be loaded slightly later.
There was a thread w/ similar error, but an eGPU that wasn't fully ready when the driver wanted to talk to it.
Do you think it would make much of a difference in this case? Seems like the 575 driver has introduced a number of issues that manifest regardless?
Offline
https://bbs.archlinux.org/viewtopic.php … 4#p2245864 - though in that case the GSP would indeed not matter.
You could check whether your specific problems are just a race condition, but yeah - the 575xx driver has some serious issues.
Online
https://bbs.archlinux.org/viewtopic.php … 4#p2245864 - though in that case the GSP would indeed not matter.
You could check whether your specific problems are just a race condition, but yeah - the 575xx driver has some serious issues.
After reading this I'm not sure I can actually fix it anymore: https://docs.nvidia.com/datacenter/tesl … own-issues
My old 1080 Ti is on VBIOS 86.02.39.00.23 and I don't think I can update it..
Offline
yeah - the 575xx driver has some serious issues.
My old 1080 Ti is on VBIOS 86.02.39.00.23 and I don't think I can update it..
You can't. And you don't need nor want to.
Lots of FUD being thrown around in various threads.
I have the same GP107 (1050 Ti) gen card as you and zero issue.
Did you get yours straightened out? I can't even tell from this mangled thread with so many tangents.
Last edited by tekstryder (2025-06-12 13:47:49)
Offline