Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.user > #240144 > unrolled thread

Debian 11: Nvidia NVS 310 with nvidia driver freezes after two days

Back to article view | Back to linux.debian.user

Contents

  Debian 11: Nvidia NVS 310 with nvidia driver freezes after two  days Roger Price <debian@rogerprice.org> - 2021-09-19 13:30 +0200
    Re: Debian 11: Nvidia NVS 310 with nvidia driver freezes after two  days "Andrew M.A. Cater" <amacater@einval.com> - 2021-09-19 14:20 +0200
      Re: Debian 11: Nvidia NVS 310 with nvidia driver freezes after two  days Roger Price <debian@rogerprice.org> - 2021-09-21 14:00 +0200
        Re: Debian 11: Nvidia NVS 310 with nvidia driver freezes after two  days Roger Price <debian@rogerprice.org> - 2021-09-25 11:40 +0200
    Re: Debian 11: Nvidia NVS 310 with nvidia driver freezes after two  days "Alexander V. Makartsev" <avbetev@gmail.com> - 2021-09-19 15:00 +0200

#240144 — Debian 11: Nvidia NVS 310 with nvidia driver freezes after two days

My Nvidia NVS 310 card with the nvidia 390.144 driver starts off perfectly, but 
after two days freezes: no reaction to keyboard or mouse action.

After reboot, command journalctl -b -1 --no-pager | grep -i -E "(nvidia|NVRM)" 
reports

  Sep 17 10:43:25 titan kernel: nvidia: loading out-of-tree module taints kernel.
  Sep 17 10:43:25 titan kernel: nvidia: module license 'NVIDIA' taints kernel.
  Sep 17 10:43:25 titan kernel: nvidia: module verification failed: signature and/or required key missing - tainting kernel
  Sep 17 10:43:25 titan kernel: nvidia-nvlink: Nvlink Core is being initialized, major device number 246
  Sep 17 10:43:25 titan kernel: nvidia 0000:03:00.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=none:owns=io+mem
  Sep 17 10:43:25 titan kernel: NVRM: loading NVIDIA UNIX x86_64 Kernel Module 390.144  Wed Jun  2 23:02:41 UTC 2021 (using threaded interrupts)
  Sep 17 10:43:25 titan kernel: nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms  390.144  Wed Jun  2 22:58:46 UTC 2021
  Sep 17 10:43:25 titan kernel: [drm] [nvidia-drm] [GPU ID 0x00000300] Loading driver
  Sep 17 10:43:25 titan kernel: [drm] Initialized nvidia-drm 0.0.0 20160202 for 0000:03:00.0 on minor 0
  Sep 17 10:43:25 titan systemd-modules-load[308]: Inserted module 'nvidia_drm'
  Sep 17 10:43:25 titan kernel: input: HDA NVidia HDMI/DP,pcm=3 as /devices/pci0000:00/0000:00:03.0/0000:03:00.1/sound/card1/input4
  Sep 17 10:43:25 titan kernel: input: HDA NVidia HDMI/DP,pcm=7 as /devices/pci0000:00/0000:00:03.0/0000:03:00.1/sound/card1/input5
  Sep 17 10:44:13 titan kernel: caller _nv027867rm+0x58/0x90 [nvidia] mapping multiple BARs
  Sep 17 10:44:13 titan kernel: caller _nv001015rm+0x1bf/0x1f0 [nvidia] mapping multiple BARs
  Sep 17 10:44:13 titan kernel: caller _nv001015rm+0x1bf/0x1f0 [nvidia] mapping multiple BARs
  Sep 19 12:04:04 titan kernel: NVRM: GPU at PCI:0000:03:00:  GPU-b2e8a5f5-e5f3-195f-0aa3-869e3fc027ac
  Sep 19 12:04:04 titan kernel: NVRM: Xid (PCI:0000:03:00): 8, Channel 00000006
  Sep 19 12:14:19 titan kernel: nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000907c:1:0
  Sep 19 12:14:21 titan kernel: nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000907c:1:0
  Sep 19 12:14:23 titan kernel: nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000907c:1:0
  Sep 19 12:14:25 titan kernel: nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000907c:1:0
  Sep 19 12:14:27 titan kernel: nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000907d:0:0
  Sep 19 12:14:29 titan kernel: nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000907c:0:0
  Sep 19 12:14:31 titan kernel: nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000907c:1:0
  Sep 19 12:14:39 titan kernel: nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000907d:0:0

inxi -SxGay reports:

System:
   Host: titan Kernel: 5.10.0-8-amd64 x86_64 bits: 64 compiler: gcc v: 10.2.1
   parameters: BOOT_IMAGE=/boot/vmlinuz-5.10.0-8-amd64
   root=UUID=faab0405-0274-4392-bca4-0f6fb6635a87 ro log_buf_len=1M
   ipv6.disable=1 net.ifnames=0 3 quiet
   Desktop: Xfce 4.16.0 tk: Gtk 3.24.24 info: xfce4-panel wm: xfwm4
   dm: LightDM 1.26.0 Distro: Debian GNU/Linux 11 (bullseye) 
Graphics:
   Device-1: NVIDIA GF119 [NVS 310] vendor: Hewlett-Packard driver: nvidia
   v: 390.144 bus ID: 03:00.0 chip ID: 10de:107d class ID: 0300
   Display: server: X.Org 1.20.11 driver: loaded: nvidia
   unloaded: fbdev,modesetting,nouveau,vesa alternate: nv
   display ID: localhost:10.0 screens: 1
   Screen-1: 0 s-res: 2880x2560 s-dpi: 96 s-size: 762x677mm (30.0x26.7")
   s-diag: 1019mm (40.1")
   Monitor-1: DP-2 res: 1440x2560 hz: 60
   Monitor-2: DP-3 res: 1440x2560 hz: 60
   OpenGL: renderer: NVS 310/PCIe/SSE2 v: 4.6.0 NVIDIA 390.144
   direct render: Yes

I still have nouveau present.  dpkg-query -l | grep nouveau reports:
  ii  libdrm-nouveau2:amd64      2.4.104-1  amd64 Userspace interface to nouveau-specific kernel DRM services -- runtime
  ii  xserver-xorg-video-nouveau 1:1.0.17-1 amd64 X.Org X server -- Nouveau display driver

Any suggestion or hint for avoiding this freeze would be much appreciated.

Roger

[toc] | [next] | [standalone]

#240147

On Sun, Sep 19, 2021 at 01:22:30PM +0200, Roger Price wrote:
> My Nvidia NVS 310 card with the nvidia 390.144 driver starts off perfectly,
> but after two days freezes: no reaction to keyboard or mouse action.
> 
> After reboot, command journalctl -b -1 --no-pager | grep -i -E
> "(nvidia|NVRM)" reports
> 
>  Sep 17 10:43:25 titan kernel: nvidia: loading out-of-tree module taints kernel.
>  Sep 17 10:43:25 titan kernel: nvidia: module license 'NVIDIA' taints kernel.
>  Sep 17 10:43:25 titan kernel: nvidia: module verification failed: signature and/or required key missing - tainting kernel
>  Sep 17 10:43:25 titan kernel: nvidia-nvlink: Nvlink Core is being initialized, major device number 246
>  Sep 17 10:43:25 titan kernel: nvidia 0000:03:00.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=none:owns=io+mem
>  Sep 17 10:43:25 titan kernel: NVRM: loading NVIDIA UNIX x86_64 Kernel Module 390.144  Wed Jun  2 23:02:41 UTC 2021 (using threaded interrupts)
>  Sep 17 10:43:25 titan kernel: nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms  390.144  Wed Jun  2 22:58:46 UTC 2021
>  Sep 17 10:43:25 titan kernel: [drm] [nvidia-drm] [GPU ID 0x00000300] Loading driver
>  Sep 17 10:43:25 titan kernel: [drm] Initialized nvidia-drm 0.0.0 20160202 for 0000:03:00.0 on minor 0
>  Sep 17 10:43:25 titan systemd-modules-load[308]: Inserted module 'nvidia_drm'
>  Sep 17 10:43:25 titan kernel: input: HDA NVidia HDMI/DP,pcm=3 as /devices/pci0000:00/0000:00:03.0/0000:03:00.1/sound/card1/input4
>  Sep 17 10:43:25 titan kernel: input: HDA NVidia HDMI/DP,pcm=7 as /devices/pci0000:00/0000:00:03.0/0000:03:00.1/sound/card1/input5
>  Sep 17 10:44:13 titan kernel: caller _nv027867rm+0x58/0x90 [nvidia] mapping multiple BARs
>  Sep 17 10:44:13 titan kernel: caller _nv001015rm+0x1bf/0x1f0 [nvidia] mapping multiple BARs
>  Sep 17 10:44:13 titan kernel: caller _nv001015rm+0x1bf/0x1f0 [nvidia] mapping multiple BARs
>  Sep 19 12:04:04 titan kernel: NVRM: GPU at PCI:0000:03:00:  GPU-b2e8a5f5-e5f3-195f-0aa3-869e3fc027ac
>  Sep 19 12:04:04 titan kernel: NVRM: Xid (PCI:0000:03:00): 8, Channel 00000006
>  Sep 19 12:14:19 titan kernel: nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000907c:1:0
>  Sep 19 12:14:21 titan kernel: nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000907c:1:0
>  Sep 19 12:14:23 titan kernel: nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000907c:1:0
>  Sep 19 12:14:25 titan kernel: nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000907c:1:0
>  Sep 19 12:14:27 titan kernel: nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000907d:0:0
>  Sep 19 12:14:29 titan kernel: nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000907c:0:0
>  Sep 19 12:14:31 titan kernel: nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000907c:1:0
>  Sep 19 12:14:39 titan kernel: nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000907d:0:0
> 
> inxi -SxGay reports:
> 
> System:
>   Host: titan Kernel: 5.10.0-8-amd64 x86_64 bits: 64 compiler: gcc v: 10.2.1
>   parameters: BOOT_IMAGE=/boot/vmlinuz-5.10.0-8-amd64
>   root=UUID=faab0405-0274-4392-bca4-0f6fb6635a87 ro log_buf_len=1M
>   ipv6.disable=1 net.ifnames=0 3 quiet
>   Desktop: Xfce 4.16.0 tk: Gtk 3.24.24 info: xfce4-panel wm: xfwm4
>   dm: LightDM 1.26.0 Distro: Debian GNU/Linux 11 (bullseye) Graphics:
>   Device-1: NVIDIA GF119 [NVS 310] vendor: Hewlett-Packard driver: nvidia
>   v: 390.144 bus ID: 03:00.0 chip ID: 10de:107d class ID: 0300
>   Display: server: X.Org 1.20.11 driver: loaded: nvidia
>   unloaded: fbdev,modesetting,nouveau,vesa alternate: nv
>   display ID: localhost:10.0 screens: 1
>   Screen-1: 0 s-res: 2880x2560 s-dpi: 96 s-size: 762x677mm (30.0x26.7")
>   s-diag: 1019mm (40.1")
>   Monitor-1: DP-2 res: 1440x2560 hz: 60
>   Monitor-2: DP-3 res: 1440x2560 hz: 60
>   OpenGL: renderer: NVS 310/PCIe/SSE2 v: 4.6.0 NVIDIA 390.144
>   direct render: Yes
> 
> I still have nouveau present.  dpkg-query -l | grep nouveau reports:
>  ii  libdrm-nouveau2:amd64      2.4.104-1  amd64 Userspace interface to nouveau-specific kernel DRM services -- runtime
>  ii  xserver-xorg-video-nouveau 1:1.0.17-1 amd64 X.Org X server -- Nouveau display driver
> 
> Any suggestion or hint for avoiding this freeze would be much appreciated.
> 
> Roger
> 

This comes down, perhaps, to having both nouveau and nididia drivers
on the same system.

I've previously had success only by installing a minimal (command line)
Debian, installing the prerequisites needed to build the proprietary Nvidia
driver, building it and only then installing X Windows / Wayland and a 
desktop environment via tasksel.

That more or less ensures that you never load nouveau at all - blacklisting
it / unloading it once installed is tedious. The merits of the proprietary
driver are that it mostly works: the downside is that you can't fix it
when something is wrong. Cue hundreds of hours of flamewars as to why
you should / shouldn't buy/use Nvidia, which drivers you should install ...

All the very best, as ever,

Andy Cater

[toc] | [prev] | [next] | [standalone]

#240219

On Sun, 19 Sep 2021, Andrew M.A. Cater wrote:

> On Sun, Sep 19, 2021 at 01:22:30PM +0200, Roger Price wrote:
>> My Nvidia NVS 310 card with the nvidia 390.144 driver starts off perfectly,
>> but after two days freezes: no reaction to keyboard or mouse action.
>>
> This comes down, perhaps, to having both nouveau and nididia drivers
> on the same system.

I use synaptic, which wants to remove far too many packages if I remove nouveau
so I didn't insist.

On Sun, 19 Sep 2021, Alexander V. Makartsev wrote:

> It looks like a hardware problem to me, even if you say it takes two days to 
> freeze. Can you tell us more about your system. Is it laptop or is it 
> stationary workstation?

The machine is a Dell 6-Core Precision WorkStation T7500, Intel Xeon E5645, with 
Bios dated 2013. The internal temperatures are: cpu: 39.0 C mobo: 26.0 C. The 
RAM total is 47.04 GiByte. I use Xfce 4.16.0.

> Was it working just fine before you upgraded to 'bullseye' (ver 5.10 kernel)?

Ran perfectly for 2 years with opensuse 42.3, a Quadro 4000 card and the nvidia 
384.69 driver.  Rock solid.  Quadro 4000 temperature typically 85C.

> Have you tried to run some benchmarks to force the issue? By doing that you 
> could reveal some potential problem with inadequate cooling or problems of...

Here is a short summary of my notes following my attempts to find a working 
setup on this workstation

     Driver: nouveau
     ---------------

Card Quadro 4000, GF100GL.  Freezes after 11 minutes with monitors lit. 
Firmware issue: failed to load nvc0_fuc084.  After freeze, journalctl shows

  fifo: INTR 0100 0000: 0...05   many repetitions
  fifo: INTR 0080 0000           once

Card NVS 310.  Freezes after 6 minutes with monitors lit. Card temperature 46C.

Card Quadro P400, GP107GL. Left monitor spontaneously rotates after 30 min. 
Other random reconfigurations.  Firmware issue: gp107/nvdec/scrubber and 
acr/DL.bin not loaded. Message: Failed to create kernel channel -22. Card 
temperature 45C.

     Nvidia drivers
     --------------

Card Quadro 4000, GF100GL. 390.144.  Freezes with blank monitors after 15 
minutes. Card temperature 85C.

Card NVS 310. 390.144.  Freezes with monitors lit after 15 mins - 3 hours.

Card Quadro P400, GP107GL. 460.91.  Card temperature 46C.  Freezes with blank 
monitors before 30 mins.

     Current situation
     -----------------

I wondered if a common feature of all the freezing was the automatic screen 
saver failing, so I installed xscreensaver and configured it to start saving my 
screen after 10 mins inactivity with the Quadro P400 card + 490.91 driver. 
This has so far held up for 22 hours.  If it holds up for a week, I will report 
it as a candidate workaround.  If it doesn't, I will still be looking for a 
solution.

Roger

[toc] | [prev] | [next] | [standalone]

#240339

On Tue, 21 Sep 2021, Roger Price wrote:

>    Nvidia drivers
>    --------------
>
> Card Quadro 4000, GF100GL. 390.144.  Freezes with blank monitors after 15 
> minutes. Card temperature 85C.
>
> Card NVS 310. 390.144.  Freezes with monitors lit after 15 mins - 3 hours.
>
> Card Quadro P400, GP107GL. 460.91.  Card temperature 46C.  Freezes with blank 
> monitors before 30 mins.
>
>    Current situation
>    -----------------
>
> I wondered if a common feature of all the freezing was the automatic screen 
> saver failing, so I installed xscreensaver and configured it to start saving 
> my screen after 10 mins inactivity with the Quadro P400 card + 490.91 driver. 
> This has so far held up for 22 hours.  If it holds up for a week, I will 
> report it as a candidate workaround.

The workaround has held up for 5 days.  I use xscreensaver options "Only One 
Screen Saver", "SlideScreen", Blank after 10 minutes, No lock screen, No Display 
Power Management.  This keeps the monitors busy.

I tried turning on Display Power Management and ended up with frozen 
monitors.

It looks as if the nvidia driver 460.91 (and other versions) breaks when 
attempting to save energy with iiyama 2783QSU monitors.

Roger

[toc] | [prev] | [next] | [standalone]

#240153

[Multipart message — attachments visible in raw view] — view raw

On 19.09.2021 16:22, Roger Price wrote:
> My Nvidia NVS 310 card with the nvidia 390.144 driver starts off 
> perfectly, but after two days freezes: no reaction to keyboard or 
> mouse action.
>
Have you tried to run some benchmarks to force the issue? By doing that 
you could reveal some potential problem with inadequate cooling or 
problems of electrical nature.
It is quite old hardware so it is hard to tell for sure. There could be 
myriad reasons why it freezes, ranging from faulty capacitors on 
motherboard and VGA to a faulty PSU.

> I still have nouveau present.  dpkg-query -l | grep nouveau reports:
>  ii  libdrm-nouveau2:amd64      2.4.104-1  amd64 Userspace interface 
> to nouveau-specific kernel DRM services -- runtime
>  ii  xserver-xorg-video-nouveau 1:1.0.17-1 amd64 X.Org X server -- 
> Nouveau display driver
>
Doesn't matter if you have 'nouveau' installed, since proprietary nvidia 
driver blacklists it for you upon installation.

> Any suggestion or hint for avoiding this freeze would be much 
> appreciated.
>
It looks like a hardware problem to me, even if you say it takes two 
days to freeze.
Can you tell us more about your system. Is it laptop or is it stationary 
workstation?
Was it working just fine before you upgraded to 'bullseye' (ver 5.10 
kernel)?
Have you tried to force reinstall 'nvidia-legacy-390xx-kernel-dkms' 
package after kernel was updated to ver 5.10?
It is a good practice to rebuild dkms module after major kernel version 
upgrade using new kernel source files.

I'd test the VGA with live Linux OS or with test Windows installation, 
just to out-rule the possible hardware problems.

-- 
With kindest regards, Alexander.

⢀⣴⠾⠻⢶⣦⠀
⣾⠁⢠⠒⠀⣿⡁ Debian - The universal operating system
⢿⡄⠘⠷⠚⠋⠀ https://www.debian.org
⠈⠳⣄⠀⠀⠀⠀

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.user

csiph-web