Wireshark · Ethereal-dev: Re: [Ethereal-dev] Segmentation fault, core dumped ethereal

[Guy Harris <guy@xxxxxxxxxx>]

> I have the following problem with Seg faults in Ethereal.  After I stop
> captures, ethereal loads the packets and promptly seg faults to a core
> dump before displaying them.

That sounds like a bug in a dissector.

The stack trace

#0  0x814df8b in filter_dialog_cb ()
#1  0x81df6ce in ypxfr1_proc ()
#2  0x814d30e in filter_dialog_cb ()
#3  0x400f68b3 in gtk_marshal_NONE__NONE () from /usr/lib/libgtk-1.2.so.0
#4  0x40128c0a in gtk_handlers_run () from /usr/lib/libgtk-1.2.so.0

	...

looks a bit bogus, given that "ypxfr1_proc" is an array of data, not a
routine, so I wouldn't infer from the core dump that it actually crashed
in "filter_dialog_cb()".

You may still have the capture file, in "/tmp" or "/var/tmp"; it'll have
a name beginning with "ether" and ending in a much of letters and
numbers.  Try reading that file in with "ethereal -r", and see if it
crashes; if it does, you might want to try running Ethereal inside GDB
and have it read that file, to see if you get a better stack trace, or
possibly send the capture file to the list or to somebody on the list
(note that if it's a large file, you might not want to send it to the
list, and there's the risk that it'd contain data that shouldn't be made
public).