The second problem with MEGACO I've found is here. It is an exception
raising when MEGACO dissector parses audit packets.
Sample audit packet looks like this:
MEGACO/1 <company2.com>:2944
Transaction=6438{
Context=*{
AuditValue=user_6{Audit{Media}}
}
}
The dissector encounters "Audit" descriptor (dissect_megaco_auditdescriptor
function) and calls dissect_megaco_descriptors() from it.
dissect_megaco_descriptors(), in turn, detects word "Media" and passes
control to dissect_megaco_mediadescriptor().
What happens inside dissect_megaco_mediadescriptor() ? The dissector tries
to find "=" character after the word "Media" which is incorrect in this
case.
I've made "quick-n-dirty" fix avoiding tvb_format_text() call if length
parameter less or equal than zero.
Please tell me am I right or wrong with it. I did not committed it to
repository because I'm not sure it is correct way to solve the problem.
Index: packet-megaco.c
===================================================================
--- packet-megaco.c (revision 18189)
+++ packet-megaco.c (working copy)
@@ -1346,8 +1346,9 @@
tvb_current_offset = tvb_find_guint8(tvb, tvb_previous_offset,
tvb_RBRKT, '=');
tokenlen = tvb_current_offset - tvb_previous_offset -1;
- proto_tree_add_text(megaco_mediadescriptor_tree, tvb,
tvb_previous_offset, tokenlen,
- "%s", tvb_format_text(tvb, tvb_previous_offset,
tokenlen));
+ if (tokenlen > 0)
+ proto_tree_add_text(megaco_mediadescriptor_tree, tvb,
tvb_previous_offset, tokenlen,
+ "%s", tvb_format_text(tvb, tvb_previous_offset,
tokenlen));
tvb_next_offset = tvb_find_guint8(tvb, tvb_previous_offset,
tvb_RBRKT, '{');
/* If a StreamID is present */
Eugene Tarlovskij
_______________________________________________
Ethereal-dev mailing list
Ethereal-dev@xxxxxxxxxxxx
http://www.ethereal.com/mailman/listinfo/ethereal-dev
