367.35 Driver - Fedora 24 GLX Segmentation Faults

My Fedora 24 systems have updated to the latest version of the NVIDIA driver (367.35) and following this I am having major issues launching a number of X applications - many commonly used (and essential) utilities do not start at all, giving a segmentation fault immediately.

This includes all the KDE apps using QT5 and GNOME apps using GTK. KDE error outputs look like:

kf5.kcoreaddons.kaboutdata: Could not initialize the equivalent properties of Q*Application: no instance (yet) existing.
failed to get the current screen resources
QXcbConnection: XCB error: 172 (Unknown), sequence: 169, resource id: 211, major code: 149 (Unknown), minor code: 20
KCrash: Application 'konsole' crashing...
KCrash: Attempting to start /usr/libexec/drkonqi from kdeinit

Even simpler X application crash, such as Emacs which fails with:

Fatal error 11: Segmentation fault
Segmentation fault (core dumped)

glxinfo and glxgears also produce segmentation faults (no other messages)

It appears as if the error is being caused by libGLX or libGLX_indirect. Uninstalling the NVIDIA driver makes the problem go away.

Is anyone else encountering this issue with the new version of the drivers and is there a fix?

Thanks for any help you can give!

Fedora 24 x86_64 here with updates-testing enabled and I’m not seeing this issue with 367.35. Using Plasma desktop fine with pleny of Qt5 apps. I install the drivers with NVIDIA’s installer - are you using the RPMFusion/unitedrpms packages?

I am using the RPM drivers from the repo http://negativo17.org/nvidia-driver/ (they are, I believe, the same as those from RPMFusion except that they include CUDA). I have found these to be more robust in the past when dealing with kernel updates, driver updates and CUDA updates and keeping everything working together without having to reinstall the drivers by hand for each new kernel.

The errors for me seem to be occurring over remote X and when remotely logged into the computer using x2go for a graphical environment. I don’t have local access to these machines to know if the error exists using the local desktop (they are in a server room).

Same error when I try to launch Emacs through x2go.
ssh -Y works though.

Personally I prefer VNC (with krfb, x0vncserver or vncserver from tigervnc package).

ssh -Y doesn’t work for me on Mac OS X client (10.11.6 using XQuartz 2.7.9)

Workaround for x2go is to uncomment (or add) the following line to /etc/x2go/x2goagent.options


This disables the GLX extension in the X server that x2go uses (in x2go the X server runs on the server and the display is forwarded to the client rather than the X server running on the client).

The workaround seems to be disabling GLX on the X server for now, though it would be good for there to be a proper fix in the NVIDIA drivers!

I have just installed the NVIDIA driver using the NVIDIA-Linux-x86_64-367.35.run from NVIDIA to a FEDORA 24 system and I am having the same issue.

For some reason, locally everything seems to be working just fine. But when I try to connect from a Mac system (with XQuartz 2.7.9) from a terminal (either with ssh -Y or ssh -XY) emacs, gedit, glxgears and several other applications give a segmentation fault error. For example :

[newton ~]$ emacs
Fatal error 11: Segmentation fault
Segmentation fault (core dumped)

Some other graphical interfaces work just fine i.e. the text editor nedit.

Would there be a workaround like disabling the GLX for ssh connections ?

Thanks a lot,

Also ran into this.

  • Remote X-server (Mageia 5, x11-server-xorg-1.16.4-2.1.mga5)
  • Host is a Fedore 23 with Dual displays
  • Running glxinfo seems enough to trigger it, but the real problem for me is gnome-terminal

gdb trace from glxinfo

Program received signal SIGSEGV, Segmentation fault.
0x00007ffff64767ad in ?? () from /lib64/libGLX_indirect.so.0
(gdb) bt
#0 0x00007ffff64767ad in ?? () from /lib64/libGLX_indirect.so.0
#1 0x00007ffff643f0b8 in ?? () from /lib64/libGLX_indirect.so.0
#2 0x00007ffff6438be1 in glXChooseVisual () from /lib64/libGLX_indirect.so.0
#3 0x0000555555556f33 in main ()