X server crashes with 319.23 and Debian 7

I upgraded to Debian Wheezy and I haven’t found a version of the driver that doesn’t crash X yet with my GeForce 9500 GT card with dual monitors attached.

When I shut down chrome (several windows and tabs), the screen sometimes pauses and flickers and then the X server shuts down.

Trying to figure out how to attach files…
nvidia-bug-report.log.gz (127 KB)
Xorg.0.log.old.txt (37.4 KB)

I forgot to mention what the crash was for the Xorg.0.log.old.txt attachment.

I pressed Ctrl-Alt-F2 to switch to a tty console and Ctrl-Alt-F7 to go back to my X display.
Then it crashed and restarted X.

I’m afraid there isn’t really enough information in your log to indicate what’s going wrong other than that the GPU is producing an error and the driver isn’t able to recover it. I’d start with checking the physical aspects of the card: is it firmly seated in the PCI-E slot, and is the heatsink clear of dust and the fan spinning smoothly? If your board has an external power connector, please make sure that’s firmly connected too (I don’t think any of these boards do, but I figured I’d ask just in case).

I haven’t been able to reproduce this problem in my testing.

I’ve been using this card for at least 2 years without incident and the only change is the new kernel and the debian upgrade.
I have turned on verbose logging for Xorg.log in /etc/gdm3/daemon.conf
Hopefully, this will reveal something in the next crash.

Not sure if related, but a number of us have been having issues with X crashes. For example, see here:


So, I haven’t had a crash lately but Chrome was constantly having errors with WebGL/OpenGL. Certain web pages would turn black and become unusable, and my .xsession-errors file in my home directory contained numerous webgl errors (and even disabled further webgl errors and told me how to turn off error reporting!)

I figured out a couple of things trying to figure out what the problem was.

Direct rendering was not working as evidenced by glxinfo | grep render

Running nvidia-installer --sanity showed that /usr/lib/libGL.la had a different checksum.

First, I uninstalled the driver in recovery mode (single-user). Then installed again.
Ran nvidia-installer --sanity and received the same error message about different checksum.

Lesson learned? Ignore the sanity check for the /usr/lib/libGL.la file

Next, I put my old /etc/X11/xorg.conf that I had saved back in place.

When I booted back into X, this time direct rendering was working! The only thing that is different in my Xorg.conf is that the Option “XAANoOffscreenPixmaps” "true’ and Option “AllowGLXWithComposite” “true” lines are in the Device section (which specifies the “nvidia” driver).

I was still having problems with misrendered characters from the font in my terminal. (see screenshot attachment)
This time, I added the Option “DamageEvents” “True” line to the Device section and that went away. It sounded like a good thing to turn on when I read the documentation.

Seems like that the config files that the NVIDIA tools create during installation are broken. So I’ll attach both NVIDIA’s and mine.

I’m disappointed that the latest non-free driver hasn’t made it into the official experimental debian repositories. Still, there are a number of bugs that I could see preventing it from being released.

Also, I can’t attach files that end in the .conf extension? Pretty lame.

xorg.conf.working.txt (2.58 KB)
xorg.conf.nvidia.txt (3.11 KB)

back to my original problem. I switched to the text console and then switched back.
Xorg had problems loading all of the video content and then crashed.

Jun 15 21:50:05 beast kernel: [ 2725.133475] NVRM: GPU at 0000:02:00: GPU-296cc2
Jun 15 21:50:51 beast kernel: [ 2771.433593] 5:3:1: cannot get freq at ep 0x84
Jun 15 21:50:51 beast kernel: [ 2771.440806] 5:3:1: cannot get freq at ep 0x84
Jun 15 21:50:51 beast kernel: [ 2771.445180] 5:3:1: cannot get freq at ep 0x84

Here’s a nice segfault/backtrace. Is anyone reading this?

Xorg.0.log.old.txt (3.63 MB)

I’m experiencing same problem on Ubuntu 13.04. Driver NVIDIA-Linux-x86_64-319.23
Who do I contact for support? I bought this expensive card and the driver crashes with segfault when starting X! None of my code is even running yet.

I have removed all configuration from /usr/share/X11/xorg.conf.d/ Only one simple xorg.conf. No display manager present.

[ 65004.065] (EE) Backtrace:
[ 65004.065] (EE) 0: /usr/bin/X (xorg_backtrace+0x36) [0x7fcf2c1dc476]
[ 65004.065] (EE) 1: /usr/bin/X (0x7fcf2c02c000+0x1b42b9) [0x7fcf2c1e02b9]
[ 65004.065] (EE) 2: /lib/x86_64-linux-gnu/libpthread.so.0 (0x7fcf2b12f000+0xfbd0) [0x7fcf2b13ebd0]
[ 65004.065] (EE) 3: /usr/lib/xorg/modules/drivers/nvidia_drv.so (0x7fcf252cd000+0x4c5623) [0x7fcf25792623]
[ 65004.065] (EE) 4: /usr/lib/xorg/modules/drivers/nvidia_drv.so (0x7fcf252cd000+0xa9e31) [0x7fcf25376e31]
[ 65004.065] (EE) 5: /usr/lib/xorg/modules/drivers/nvidia_drv.so (0x7fcf252cd000+0xaa415) [0x7fcf25377415]
[ 65004.065] (EE) 6: /usr/lib/xorg/modules/drivers/nvidia_drv.so (0x7fcf252cd000+0x645ee) [0x7fcf253315ee]
[ 65004.065] (EE) 7: /usr/lib/xorg/modules/drivers/nvidia_drv.so (0x7fcf252cd000+0xdc8da) [0x7fcf253a98da]
[ 65004.065] (EE) 8: /usr/lib/xorg/modules/drivers/nvidia_drv.so (0x7fcf252cd000+0xedc00) [0x7fcf253bac00]
[ 65004.065] (EE) 9: /usr/lib/xorg/modules/drivers/nvidia_drv.so (0x7fcf252cd000+0x81066) [0x7fcf2534e066]
[ 65004.065] (EE) 10: /usr/lib/xorg/modules/drivers/nvidia_drv.so (0x7fcf252cd000+0x81468) [0x7fcf2534e468]
[ 65004.065] (EE) 11: /usr/lib/xorg/modules/drivers/nvidia_drv.so (0x7fcf252cd000+0x4ade1d) [0x7fcf2577ae1d]
[ 65004.065] (EE) 12: /usr/lib/xorg/modules/drivers/nvidia_drv.so (0x7fcf252cd000+0x4ae025) [0x7fcf2577b025]
[ 65004.065] (EE) 13: /usr/lib/xorg/modules/drivers/nvidia_drv.so (0x7fcf252cd000+0x4a4c15) [0x7fcf25771c15]
[ 65004.065] (EE) 14: /usr/bin/X (AddScreen+0x85) [0x7fcf2c084d55]
[ 65004.065] (EE) 15: /usr/bin/X (InitOutput+0x402) [0x7fcf2c0c6662]
[ 65004.065] (EE) 16: /usr/bin/X (0x7fcf2c02c000+0x47386) [0x7fcf2c073386]
[ 65004.066] (EE) 17: /lib/x86_64-linux-gnu/libc.so.6 (__libc_start_main+0xf5) [0x7fcf29d7cea5]
[ 65004.066] (EE) 18: /usr/bin/X (0x7fcf2c02c000+0x478c1) [0x7fcf2c0738c1]
[ 65004.066] (EE) 
[ 65004.066] (EE) Segmentation fault at address 0x3f0
[ 65004.066] 
Fatal server error:
[ 65004.066] Caught signal 11 (Segmentation fault). Server aborting

another week, another crash.
nvidia-bug-report.log.gz (222 KB)
Xorg.0.log.txt (6.33 MB)

I’m now on driver 331.20 and I still have crashes.
I’m going to be getting a new video card that isn’t made by NVIDIA.

I ended up getting the VisionTek 900358 Radeon HD 5450 from newegg.

Much more stable so far.

Hope this helps!

Probably the best part is the amount of configuration cruft that I was able to remove from the xorg.conf all related to configuring the nvidia driver.

Almost as good as that is the xrandr(1) utility that I can tell the radeon card how to configure dual displays on the fly without restarting X. Pretty amazing, if you ask me.


  • Is this issue repro with single monitor connected?
  • What desktop environment you are using kde, gnome, xfce, gnome-shell, unity, etc ?
  • I think you opened many tabs on chrome browser and then did VT Switch, Is issue repro just by VT switch? Or I have to open different tabs in chrome and then VT switch? What are the all website your are accessing ? Does there web sites includes contents like adobe flash, video, html5 etc . ?