570 release feedback & discussion

Hi, recently switched to 570.86.16, kernel version 6.13.2 – driver seems to fail/crash for me now

Feb 18 21:32:31 labtop kernel: NVRM: GPU 0000:01:00.0: RmInitAdapter failed! (0x62:0xffff:2521)
Feb 18 21:32:31 labtop kernel: NVRM: GPU 0000:01:00.0: rm_init_adapter failed, device minor number 0
Feb 18 21:32:31 labtop kernel: [drm:nv_drm_load [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to allocate NvKmsKapiDevice
Feb 18 21:32:31 labtop kernel: [drm:nv_drm_register_drm_device [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to register device

nvidia-bug-report.log (1.1 MB)

Rolled back to 565.77 for now where I am not experiencing this issue

2 Likes

If the dGPU is not active at all my temps are low, between 36c and 40. As soon as steam is running in the background the temps are increasing by 10c. With driver 565 when steam is active i am around 38-42c so it did increase a bit. But so far the driver runs great.

Do you use suspend? I experience that occasionally, and since I have set fbdev to 0 more rarely I think but it could be subjective, wrong impression. I haven’t experienced issue in a while, but I think there’s a connection between suspend and resuming from suspend and this issue in my case, or it’s just random lol. It might be a coincidence, but I experience the issue always when I resume from suspend (For this I use Xorg) then for whatever reason (Usually just testing) I log into a Wayland session. One potential argument indicating it might be related to suspend (Or something else) is that when the issue starts happening, it can reliably be replicated. E.g. simply playing a video in a Firefox (I use deb version in case that it matters but people experience the issue with snap version too AFAIK) then entering full screen will cause the DE to freeze. Sound usually continues playing. Switching to one of the TTYs will often, but not always, help, but as soon as I switch the video to full screen again it will freeze again. Rebooting seems to reliably solve the issue, and it doesn’t come back—unless by coincidence—until I have resumed from suspend and logged into a Wayland session again.

The monitor freeze/flip issue is definitely not related to suspend as this is my software engineering workstation and suspend states are disabled as it is always busy doing something for work. The only power saving feature that is enabled is monitor turn off on 2 minutes after KDE lock but this has never happened while powering on the monitors. Most of the times it happened it was while playing a video in a secondary firefox window on the second monitor and a couple of times while running a game (not the one I’m working on), under Proton Experimental. My steam client is set to [bleeding edge] PE build as I’m testing my code on it most of the time and I’m too lazy to switch back and forth when running an actual game.
That said I just upgraded to kernel 6.12.15 from 6.12.13 and I see some changes in .14 changelog (.15 is just XFS hotfix) that sound like they could be related. I will report if it triggers again.

you’re right. I was digging through journalctl output last night and it appears that occasionally the driver (570) segfaults. Tho, in my case this has rarely (IIRC once, and it was yesterday) happened unless I have resumed from suspend. The issues of X ‘gracefully’ (or so I thought because there’s nothing in Xorg log files) closing, and full screen videos freezing the DE might be related, because on my machine they always happen together at least partially. By that I mean if X11 session has crashed, the video will freeze the DE in Wayland session.

I’m in a hurry so I hope I have copied the relevant part of the output (Should contain segmentation fault part):

Feb 20 00:22:50 denis-ws polkitd[1596]: Unregistered Authentication Agent for unix-session:c4 (system bus name :1.600, object path /org/freedesktop/PolicyKit1/AuthenticationAgent, locale en_US.UT>
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (II) UnloadModule: “libinput”
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (II) systemd-logind: releasing fd for 13:67
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (II) UnloadModule: “libinput”
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (II) systemd-logind: releasing fd for 13:64
Feb 20 00:22:50 denis-ws gnome-shell[54542]: DING: Gtk-Message: 00:22:50.802: Failed to load module “xapp-gtk3-module”
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (II) UnloadModule: “libinput”
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (II) systemd-logind: releasing fd for 13:66
Feb 20 00:22:50 denis-ws dbus-daemon[54175]: [session uid=1000 pid=54175] Activating via systemd: service name=‘org.gtk.vfs.Metadata’ unit=‘gvfs-metadata.service’ requested by ‘:1.96’ (uid=1000 p>
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (II) UnloadModule: “libinput”
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (II) systemd-logind: releasing fd for 13:65
Feb 20 00:22:50 denis-ws gnome-shell[54542]: DING: Detected async api for thumbnails
Feb 20 00:22:50 denis-ws systemd[54135]: Starting gvfs-metadata.service - Virtual filesystem metadata service…
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (II) NVIDIA(GPU-0): Deleting GPU-0
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (II) Server terminated successfully (0). Closing log file.
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE)
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE) Backtrace:
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE) 0: /usr/lib/xorg/Xorg (OsLookupColor+0x14c) [0x652be67a5dbc]
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE) 1: /lib/x86_64-linux-gnu/libc.so.6 (__sigaction+0x50) [0x734ab8645330]
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE) unw_get_proc_name failed: no unwind info found [-10]
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE) 2: /lib/x86_64-linux-gnu/libnvidia-glcore.so.570.86.16 (?+0x0) [0x734ab646cd62]
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE) unw_get_proc_name failed: no unwind info found [-10]
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE) 3: /lib/x86_64-linux-gnu/libGLX_nvidia.so.0 (?+0x0) [0x734ab7e9a7c7]
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE) unw_get_proc_name failed: no unwind info found [-10]
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE) 4: /lib/x86_64-linux-gnu/libGLX_nvidia.so.0 (?+0x0) [0x734ab7e9aedf]
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE) unw_get_proc_name failed: no unwind info found [-10]
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE) 5: /lib/x86_64-linux-gnu/libGLX_nvidia.so.0 (?+0x0) [0x734ab7f2b009]
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE) 6: /lib/x86_64-linux-gnu/libc.so.6 (erand48_r+0x1d6) [0x734ab8647a76]
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE) 7: /lib/x86_64-linux-gnu/libc.so.6 (exit+0x1e) [0x734ab8647bbe]
Feb 20 00:22:50 denis-ws dbus-daemon[54175]: [session uid=1000 pid=54175] Successfully activated service ‘org.gtk.vfs.Metadata’
Feb 20 00:22:50 denis-ws systemd[54135]: Started gvfs-metadata.service - Virtual filesystem metadata service.
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE) 8: /lib/x86_64-linux-gnu/libc.so.6 (__libc_init_first+0x91) [0x734ab862a1d1]
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE) 9: /lib/x86_64-linux-gnu/libc.so.6 (__libc_start_main+0x8b) [0x734ab862a28b]
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE) 10: /usr/lib/xorg/Xorg (_start+0x25) [0x652be6613395]
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE)
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE) Segmentation fault at address 0x8
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE)
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: Fatal server error:
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE) Caught signal 11 (Segmentation fault). Server aborting
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE)
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE)
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: Please consult the The X.Org Foundation support
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: at http://wiki.x.org
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: for help.
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE) Please also check the log file at “/var/lib/gdm3/.local/share/xorg/Xorg.0.log” for additional information.
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE)
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (WW) xf86CloseConsole: KDSETMODE failed: Bad file descriptor
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (WW) xf86CloseConsole: VT_GETMODE failed: Bad file descriptor
Feb 20 00:22:50 denis-ws /usr/libexec/gdm-x-session[52180]: (EE) systemd-logind: ReleaseControl failed: You are not in control of this session
Feb 20 00:22:50 denis-ws gnome-shell[54542]: DING: (gjs:55227): Gjs-WARNING **: 00:22:50.867: GLib.unix_signal_add has been moved to a separate platform-specific library. Please update your code >
Feb 20 00:22:50 denis-ws gnome-shell[54542]: DING: 0 DesktopManager() [“/usr/share/gnome-shell/extensions/ding@rastersoft.com/app/desktopManager.js”:263:12]
Feb 20 00:22:50 denis-ws gnome-shell[54542]: DING: 1 anonymous() [“/usr/share/gnome-shell/extensions/ding@rastersoft.com/app/ding.js”:180:25]
Feb 20 00:22:50 denis-ws gnome-shell[54542]: DING: 2 anonymous() [“/usr/share/gnome-shell/extensions/ding@rastersoft.com/app/ding.js”:197:20]
Feb 20 00:22:50 denis-ws gnome-shell[54542]: DING: 3 [“/usr/share/gnome-shell/extensions/ding@rastersoft.com/app/ding.js”:206:12]
Feb 20 00:22:50 denis-ws gnome-shell[54542]: DING: DBus interface for Nautilus (org.gnome.Nautilus.FileOperations2) is now available.
Feb 20 00:22:50 denis-ws gnome-shell[54542]: DING: DBus interface for Nautilus (org.freedesktop.FileManager1) is now available.
Feb 20 00:22:50 denis-ws gnome-shell[54542]: DING: DBus interface for Gvfs daemon (org.gtk.vfs.Metadata) is now available.
Feb 20 00:22:50 denis-ws gnome-shell[54542]: DING: DBus interface for Switcheroo control (net.hadess.SwitcherooControl) is now available.
Feb 20 00:22:50 denis-ws gnome-shell[54542]: DING: ** Message: 00:22:50.883: Connecting to org.freedesktop.Tracker3.Miner.Files
Feb 20 00:22:50 denis-ws gnome-shell[54542]: DING: GNOME nautilus 46.2
Feb 20 00:22:51 denis-ws gdm-launch-environment][52098]: pam_unix(gdm-launch-environment:session): session closed for user gdm
Feb 20 00:22:51 denis-ws systemd[1]: session-c4.scope: Deactivated successfully.
Feb 20 00:22:51 denis-ws systemd[1]: session-c4.scope: Consumed 4.334s CPU time.
Feb 20 00:22:51 denis-ws systemd-logind[1663]: Session c4 logged out. Waiting for processes to exit.
Feb 20 00:22:51 denis-ws systemd-logind[1663]: Removed session c4.
Feb 20 00:22:52 denis-ws kernel: [drm:nv_drm_master_set [nvidia_drm]] ERROR [nvidia-drm] [GPU ID 0x00000100] Failed to grab modeset ownership
Feb 20 00:22:52 denis-ws kernel: [drm:nv_drm_master_set [nvidia_drm]] ERROR [nvidia-drm] [GPU ID 0x00000100] Failed to grab modeset ownership
Feb 20 00:22:52 denis-ws kernel: [drm:nv_drm_master_set [nvidia_drm]] ERROR [nvidia-drm] [GPU ID 0x00000100] Failed to grab modeset ownership
Feb 20 00:22:52 denis-ws kernel: [drm:nv_drm_master_set [nvidia_drm]] ERROR [nvidia-drm] [GPU ID 0x00000100] Failed to grab modeset ownership
Feb 20 00:22:52 denis-ws kernel: [drm:nv_drm_master_set [nvidia_drm]] ERROR [nvidia-drm] [GPU ID 0x00000100] Failed to grab modeset ownership
Feb 20 00:22:52 denis-ws kernel: [drm:nv_drm_master_set [nvidia_drm]] ERROR [nvidia-drm] [GPU ID 0x00000100] Failed to grab modeset ownership
Feb 20 00:22:52 denis-ws kernel: [drm:nv_drm_master_set [nvidia_drm]] ERROR [nvidia-drm] [GPU ID 0x00000100] Failed to grab modeset ownership
Feb 20 00:22:52 denis-ws kernel: [drm:nv_drm_master_set [nvidia_drm]] ERROR [nvidia-drm] [GPU ID 0x00000100] Failed to grab modeset ownership
Feb 20 00:22:52 denis-ws gdm3[2507]: Gdm: Child process -52172 was already dead.
Feb 20 00:22:52 denis-ws kernel: [drm:nv_drm_master_set [nvidia_drm]] ERROR [nvidia-drm] [GPU ID 0x00000100] Failed to grab modeset ownership
Feb 20 00:22:52 denis-ws kernel: [drm:nv_drm_master_set [nvidia_drm]] ERROR [nvidia-drm] [GPU ID 0x00000100] Failed to grab modeset ownership
Feb 20 00:22:52 denis-ws kernel: [drm:nv_drm_master_set [nvidia_drm]] ERROR [nvidia-drm] [GPU ID 0x00000100] Failed to grab modeset ownership
Feb 20 00:22:52 denis-ws kernel: [drm:nv_drm_master_set [nvidia_drm]] ERROR [nvidia-drm] [GPU ID 0x00000100] Failed to grab modeset ownership
Feb 20 00:22:52 denis-ws gdm3[2507]: Gdm: on_display_removed: assertion ‘GDM_IS_REMOTE_DISPLAY (display)’ failed
Feb 20 00:22:52 denis-ws gnome-shell[54542]: Window manager warning: Overwriting existing binding of keysym 32 with keysym 32 (keycode b).
Feb 20 00:22:52 denis-ws gnome-shell[54542]: Window manager warning: Overwriting existing binding of keysym 33 with keysym 33 (keycode c).
Feb 20 00:22:52 denis-ws gnome-shell[54542]: Window manager warning: Overwriting existing binding of keysym 35 with keysym 35 (keycode e).
Feb 20 00:22:52 denis-ws gnome-shell[54542]: Window manager warning: Overwriting existing binding of keysym 38 with keysym 38 (keycode 11).


This came up on the Arch forum and reminded me to look into it further:

I removed the kms hook and rebuilt initramfs for all kernels. The ~50MB “bloat” is vanquished.

Rebooted into each of them to assure no shenanigans ensued. All systems nominal.

Before:

-rw-r-----  1 root root 197M Feb  3 10:37  initramfs-linux.img
-rw-r-----  1 root root 197M Feb  3 10:37  initramfs-linux-lts.img
-rw-r-----  1 root root 155M Feb  3 10:37  initramfs-linux-stable-mfm.img
-rw-r-----  1 root root 155M Feb  3 10:37  initramfs-linux-zen-mfm.img
-rw-r-----  1 root root  75M Jan 17 12:01  initrd.img-6.1.0-30-amd64

After:

-rw-r-----  1 root root 152M Feb 20 13:15  initramfs-linux.img
-rw-r-----  1 root root 152M Feb 20 13:09  initramfs-linux-lts.img
-rw-r-----  1 root root 152M Feb 20 13:19  initramfs-linux-stable-mfm.img
-rw-r-----  1 root root  75M Jan 17 12:01  initrd.img-6.1.0-30-amd64

Same issue happening with a single monitor, it keeps happening seemingly at random. It freezes, takes some time before it can get to terminal, and then can switch back and it usually works. This is new since 570, the fbdev switch ended up not helping

I noticed a lot of RIP: 0010:drm_gem_prime_fd_to_handle.cold in dmesg.
Not sure what caused it.

nvidia-bug-report.log.gz (2.5 MB)

After years and years the 0rpm mode also fails to work most of the time for me on my RTX 3070 founders edition model. This screenshot from nvtop shows that my GPU is 36c yet the fan speed is at 30%. Although 30% doesn’t sound like too much, they are easily the loudest in my chassis.

I am using the proprietary kernel modules with GSP disabled. There should be no reason for the fans to be spinning at a temperature this low.

It’s funny because sometimes I catch my GPU at the high 50’s with the fan speed on 0%. So what is going on?? Is the fan speed dependent on something else? Why would this be the case? The power usage was only 17W, and the card was in an idle state (other than the high vram issues which have STILL not been fixed)

Still no Linux stable driver for RTX 5000 users after 3 weeks ?

They released one yesterday, 20 FEB (to support 5070ti), but it seems to be a CI/CD (release) problem because it’s the same 570.86.16. (released on 30 Jan) Driver Details | NVIDIA

I’ve been having a very annoying issue since the 565 drivers, where all things that run with xwayland (games, discord, jellyfinmediaplayer) are running at horrible framerate (max i’ve seen was about 15). If i make the window smaller, the framerate does go up, but the only way i can get around this issue for now is by using gamescope for my games and just having to deal with the other apps not working properly

I’m having trouble installing 570 from the run file on a computer with a geforce 4070. I had been using 567 on Linux 6.12 and I’m trying to install 570 on 6.13 and getting an error during installation:

“nvidia_drm: Unknown symbol drm_client_setup (err -2)”

I’m using essentially the same linux kernel config settings from 6.12 to 6.13, but can’t get the nvidia driver to install on 6.13. Here’s DRM stuff from my kernel config. Am I missing something I need for 6.13?

CONFIG_DRM=y
CONFIG_DRM_MIPI_DSI=y

CONFIG_DRM_DEBUG_MM is not set

CONFIG_DRM_KMS_HELPER=m
CONFIG_DRM_PANIC=y
CONFIG_DRM_PANIC_FOREGROUND_COLOR=0xffffff
CONFIG_DRM_PANIC_BACKGROUND_COLOR=0x000000
CONFIG_DRM_PANIC_DEBUG=y
CONFIG_DRM_PANIC_SCREEN=“user”

CONFIG_DRM_DEBUG_DP_MST_TOPOLOGY_REFS is not set

CONFIG_DRM_DEBUG_MODESET_LOCK is not set

CONFIG_DRM_CLIENT=y
CONFIG_DRM_CLIENT_LIB=m
CONFIG_DRM_CLIENT_SELECTION=m
CONFIG_DRM_CLIENT_SETUP=y

CONFIG_DRM_FBDEV_EMULATION=y
CONFIG_DRM_FBDEV_OVERALLOC=100

CONFIG_DRM_FBDEV_LEAK_PHYS_SMEM is not set

CONFIG_DRM_LOAD_EDID_FIRMWARE=y
CONFIG_DRM_DISPLAY_HELPER=m

CONFIG_DRM_DISPLAY_DP_AUX_CEC is not set

CONFIG_DRM_DISPLAY_DP_AUX_CHARDEV is not set

CONFIG_DRM_DISPLAY_DP_HELPER=y
CONFIG_DRM_DISPLAY_DP_TUNNEL=y
CONFIG_DRM_DISPLAY_DSC_HELPER=y
CONFIG_DRM_DISPLAY_HDCP_HELPER=y
CONFIG_DRM_DISPLAY_HDMI_HELPER=y
CONFIG_DRM_TTM=m
CONFIG_DRM_EXEC=m
CONFIG_DRM_GPUVM=m
CONFIG_DRM_BUDDY=m
CONFIG_DRM_VRAM_HELPER=m
CONFIG_DRM_TTM_HELPER=m
CONFIG_DRM_GEM_SHMEM_HELPER=m
CONFIG_DRM_SUBALLOC_HELPER=m
CONFIG_DRM_SCHED=m

CONFIG_DRM_NOUVEAU is not set

CONFIG_FB=y
CONFIG_FB_SIMPLE=m
CONFIG_FB_CORE=y
CONFIG_FB_NOTIFY=y
CONFIG_FIRMWARE_EDID=y
CONFIG_FB_DEVICE=y

CONFIG_FB_NVIDIA is not set

1 Like

Voltage readings no longer work on 570 driver (N/A), works fine on 565.
Tested with nvidia-smi --query --display=VOLTAGE.

2 Likes

I don’t get it. I’ve seen people all over the internet complaining about that, but almost nobody bothered to check that the 4060 does not have power monitoring!!!

Right from the lions mouth

Same issue on a 4080 super which does have power monitoring. no voltages shown in nvidia-smi -q -d VOLTAGE

It does. Just not lower power usage.

How to force Ubuntu 24.10 to use Wayland instead of x11 on Nvidia Drivers 565.77 and 570.86 Beta on RTX 2070

Using NVIDIA driver (open-kernel) metapackage from nvidia-driver-570-open (proprietary) gives you Wayland on Ubuntu 24.10 with the PPA installation method! 🧐

570.86.16 and 550.144.03
OpenGL still broken. Can’t Nvidia suspend everything else and finally fix critical issues?