Wireshark-bugs: [Wireshark-bugs] [Bug 8518] New: glibc detected /usr/bin/tshark: double free or

Date: Sun, 24 Mar 2013 22:57:54 +0000
Bug ID 8518
Summary glibc detected /usr/bin/tshark: double free or corruption when using "-r -"
Classification Unclassified
Product Wireshark
Version SVN
Hardware x86-64
OS Mac OS X 10.8
Status UNCONFIRMED
Severity Normal
Priority Low
Component TShark
Assignee [email protected]
Reporter [email protected]

Created attachment 10450 [details]
The capture that causes the crash.

Build Information:
Latest version from the git repo (assuming it's the same as SVN trunk)

Also on both latest OS X and Debian.
--
Certain capture files trigger this segfault where it only happens when the
capture file is being streamed into tshark. e.g.:

# ./tshark -r ~/Downloads/net-2009-12-04-11_47.dmp
  1   0.000000 SmcNetwo_81:db:10 -> LLDP_Multicast LLDP 118 Chassis Id =
00:22:2d:81:db:10 Port Id = 1 TTL = 120

# cat ~/Downloads/net-2009-12-04-11_47.dmp | ./tshark -r -
tshark(35000) malloc: *** error for object 0x7fd0923475f0: pointer being freed
was not allocated
Abort trap: 6

Steps to debug:

# mkfifo capture
# cat ~/Downloads/net-2009-12-04-11_47.dmp > capture &
# gdb ./tshark
(gdb) run -r - <capture

Backtrace:

Program received signal SIGABRT, Aborted.
0x00007fff8bdafd46 in __kill ()
(gdb) bt
#0  0x00007fff8bdafd46 in __kill ()
#1  0x00007fff8bc9fdf0 in abort ()
#2  0x00007fff8bc739b9 in free ()
#3  0x0000000100090d6f in wtap_close ()
#4  0x000000010006fe26 in wtap_open_offline ()
#5  0x00000001000106d5 in cf_open (cf=0x100039900, fname=0x1065be320 "-",
is_tempfile=0, err=0x7fff5fbfe764) at tshark.c:3647
#6  0x000000010000f188 in main (argc=1606413808, argv=0x7fff5fbfedf0) at
tshark.c:1848

>From my initial debugging it happens at around file_access.c:366 where it is
failing on the seek:

if (file_seek(wth->fh, 0, SEEK_SET, err) == -1) {
            /* I/O error - give up */
            wtap_close(wth);
            return NULL;
        }

Unfortunately my budget with tshark has run out at this stage, so I didn't have
time to find the problem and make a patch. Hopefully this report is useful.


You are receiving this mail because:
  • You are watching all bug changes.