https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=3395
--- Comment #10 from Bill Meier <wmeier@xxxxxxxxxxx> 2009-07-27 16:43:21 PDT ---
(In reply to comment #7)
> Created an attachment (id=3203)
--> (https://bugs.wireshark.org/bugzilla/attachment.cgi?id=3203) [details]
> Example capture 2
>
Thanks for taking the time and effort to add defragmentation code to the NDMP
dissector and for uploading the capture files.
It turned out that the defragmentation didn't quite work when the TCP
preference "Try heuristic sub-dissectors first" was off (the default).
See Bug #3755.
I've committed a fix for that problem (and one or two others) in SVN #29208
(http://anonsvn.wireshark.org/viewvc?view=rev&revision=29208).
Please let me know if you have any comments.
A question: Is it the case for NDMP that for a particular conversation only one
sequence of fragments of a PDU can be "in flight" at any one time ?
If so, I suspect the defragmentation code can be simplified by just using a
"per conversation" "id" (eg: conversation->index) rather than using the start
tcp sequence number of each fragment group as the id.
Doing this, along with using fragment_add_seq_next() rather than
fragment_add_seq_check() would seem to do away with the need for the fragsA and
fragsB hash & etc.
As far as I know, there is no need for each set of fragments for a conversation
have a distinct id as long as the fragment sets don't overlap. [If I'm wrong
I'm sure someone will jump in with a correction].
--
Configure bugmail: https://bugs.wireshark.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.
