Nvidia driver intermittent IRQ error and freeze on boot.

English Hardware
1

I get a freeze during boot that I think is caused by this:

    30.813] (EE) NVIDIA(GPU-0): The NVIDIA kernel module does not appear to be receiving
    30.813] (EE) NVIDIA(GPU-0):     interrupts generated by the NVIDIA GPU at PCI:2:0:0. 
    30.813] (EE) NVIDIA(GPU-0):     Please see Chapter 8: Common Problems in the README for
    30.813] (EE) NVIDIA(GPU-0):     additional information.

The freeze gives me a blank screen with cursor but completely unresponsive to anything other than hard reboot.

The readme (here https://download.nvidia.com/XFree86/Linux-x86/384.59/README/commonproblems.html ) says its IRQ relate but when I look for one with lsdev it seems fine - but then I can only run lsdev when its booted.

How can I find the cause of what seems to be an intermitted IRQ problem given it stops booting?

Thanks for any help. I cannot figure this out at all.

2

Please provide more info.
Does it works before?

Post


inxi -Gxx
3

The output if inxi is

Graphics:  Card: NVIDIA GK107GL [Quadro K600] bus-ID: 02:00.0 chip-ID: 10de:0ffa
           Display Server: x11 (X.Org 1.20.3 ) drivers: nvidia (unloaded: modesetting,fbdev,vesa,nouveau)
           Resolution: 1280x1024@60.02hz
           OpenGL: renderer: Quadro K600/PCIe/SSE2
           version: 4.5.0 NVIDIA 430.34 (compat-v: 4.6.0) Direct Render: Yes

I have not had the problem in the last few days so its possible that a recent driver update date had fixed it. I usually only reboot once a day so it could just be luck.

4

It has started happening again.

Same message in the logs. Still intermittent. I have enabled SysReq keys so I can see if that gives me a cleaner reboot.

5

I would:
Delete all nvidia-driver-pacckages from Repo, disable the Repo, download the latest nvidia-driver from here:
http://download.nvidia.com/XFree86/Linux-x86_64/
make the driver executable (chmod +x), boot to Runlevel 3 (multiuser.target) and install the downloaded driver.

6

I have got a few upgrades since the above, and the driver is the same version as the latest on the download page. Is it still worth trying installing manually?

7

Are you using AMD hardware?
Possible issues with a IOMMU.

Changing videocard can easy solve your problem.

8

Ty - In BIOS change iommu to software rather then hardware. This helped my iommu problems.

9

It turned out to be pretty much that, but not quite.

I had disabled interrupt remapping (VT-D) because I knew there were issue with it, but the driver picked up that there were problems and worked around them.

Re-enabling it made the problems go away.

I do get an occasional “graphics restart” or two in KDE after wake from suspend, but it recovers in seconds so is just an irritant (especially as the machine is a desktop).

Thanks for help.