Wireshark-bugs: [Wireshark-bugs] [Bug 1076] RTP analysis' jitter values are thrown off by RTP events

From: bugzilla-daemon@xxxxxxxxxxxxx
Date: Tue, 4 Sep 2007 01:01:54 +0000 (GMT)

http://bugs.wireshark.org/bugzilla/show_bug.cgi?id=1076


smaximov@xxxxxxxx changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
    Attachment #879|                            |review_for_checkin?
               Flag|                            |




------- Comment #3 from smaximov@xxxxxxxx  2007-09-04 01:01 GMT -------
Created an attachment (id=879)
 --> (http://bugs.wireshark.org/bugzilla/attachment.cgi?id=879&action=view)
fix for dynamic payload dissection and jitter computation

The original bug 1076 was rooted in a number of problems:

a) Ephemeral string (packet duration) was inserted into long-term hash of
dynamic payloads.

The memory would be rewritten for each new input packet resulting in trashed
value strings in the hash of dynamic payloads rtp_dyn_payload. As a solution,
g_strdup() is used for hash insertion. To guarantee proper release of hash
values, hash is now created with g_hash_table_new_full() passing g_free for
both, key and value of the hash. This will also eliminate minor memory leak due
to keys not being freed.

With this fix, when time comes to lookup MIME type of dynamic payload, the MIME
string is not corrupted anymore.

a) There was no clock_rate mapping for MIME type "telephone-event". 

Lack of default clock rate mapping would result in clock_rate being set to 1.
This would result in huge jump in jitter estimate. Added definition of default
clock rate for dynamic payload "telephone-even" as per RFC4733.


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