Wireshark-bugs: [Wireshark-bugs] [Bug 2470] Buildbot crash output: fuzz-2008-04-16-7479.pcap

Date: Wed, 21 May 2008 11:05:22 -0700 (PDT)
http://bugs.wireshark.org/bugzilla/show_bug.cgi?id=2470


Bill Meier <wmeier@xxxxxxxxxxx> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED




--- Comment #6 from Bill Meier <wmeier@xxxxxxxxxxx>  2008-05-21 11:05:19 PDT ---
Fixed in SVN #25343.

Essentially: An attempt to reassemble a sequence of frames like the following
caused a crash in fragment_add_work() in reassemble.c

(The offsets, etc values are from the extract pcap attached).

frame #1:  offset:     0   len: 1460 
frame #2:  offset:  1460   len: 1460
frame #3:  offset:  2920   len: 1460
frame #4:  offset: 15663   len: 1460
frame #5:  offset: 15663   len: 1460
frame #6:  offset:  2299   len: 1460;  Flags: FIN + ...

Notes:

'DESEGMENT_UNTIL_FIN' was involved requiring a reassembly 
of the fragments in the 6 frames;

The crash (which did not consistently occur on Windows) was due to a 
memcmp() in fragment_add_work comparing beyond the end of a g_malloc'd buffer.

I'm guessing that the occurrence of the crash depended upon the 
exact sequence of memory allocations, etc.

In fact, for tshark the crash would occur 
or not occur depending upon the options supplied to tshark. (-n, etc)


-- 
Configure bugmail: http://bugs.wireshark.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.