I have a fedora 13 x64 box I use at home for a storage server. I use sshfs to mount my music from home and listen at work, I run my irssi client in screen on it for freenode communication and lately I appears to be popping up with protection faults from the kernel.

The message I receive:

Segmentation fault (core dumped)
[sanchez@media ~]$
Message from syslogd@media at Oct  8 08:42:00 ...
 kernel:general protection fault: 0000 [#3] SMP

Message from syslogd@media at Oct  8 08:42:00 ...
 kernel:last sysfs file: /sys/devices/pci0000:00/0000:00:1d.2/usb8/devnum

Message from syslogd@media at Oct  8 08:42:00 ...

Message from syslogd@media at Oct  8 08:42:00 ...
 kernel:Call Trace:

Message from syslogd@media at Oct  8 08:42:00 ...
 kernel:Code: c9 74 5b 66 8b 56 04 66 8b 4e 02 66 8b 36 44 0f b7 d1 0f b7 c6 c1 
e0 09 42 8d 04 90 8d 04 02 23 47 10 0f b7 c0 49 8b 04 c1 eb 2c <66> 3b 30 75 21 
66 8b 78 02 66 39 f9 75 0d 66 3b 50 04 75 07 66

My machine is:

Linux #1 SMP Wed Sep 15 03:36:55 UTC 2010 x86_64 GNU/Linux

It appears to be related to a usb device, because of the last sysfs file: /sys/devices/pci0000:00/0000:00.1d.2/usb8/devnum however I am not using an usb ports it is a headless desktop with all internal hard drives and a ethernet running to my switch.

Any help on this would be appreciated?

  • 6
    Test system with memtest86+. Does it report any problems? Also - did it worked before with Or any other kernel? Oct 8, 2010 at 13:00
  • I just got into work and was able to ssh to the machine, this machine has been running for 6+ months and this just started happening in the last week. It would be ashame if its a memory issue as I have 2x2gb of crucial ballistix (its an old gaming rig) and they were replaced when I rebuilt the system about 6 months ago because they were defective. Will run memtest this evening/weekend when I get home to the box.
    – Chris
    Oct 8, 2010 at 13:06
  • 1
    Could also be bad disk, but RAM is more likely. Power supply could also cause problems.
    – Wodin
    Jan 26, 2011 at 7:25
  • RAM is something that's not hard to test for, and memory issues (which include incompatibilities, not only bad modules) can cause so many complex and different effects that, no matter what you're testing for, if you don't have a clear hint, you should run a memory test.
    – njsg
    Feb 13, 2012 at 14:06
  • We never heard back about memtest, so I'm going to close this Feb 13, 2012 at 15:28

1 Answer 1


You can examine core files with gdb. However, in this case, you seem to have several consecutive memory faults, and a good advise would be to check your RAM using memory diagnostic programs. Those are often found on distribution's liveCDs.

Not the answer you're looking for? Browse other questions tagged .