Wireshark mailing list archives

Re: Handling TCP packets reordering


From: Sake Blok <sake () euronet nl>
Date: Wed, 4 May 2011 23:29:51 +0200

On 4 mei 2011, at 22:48, Jeff Morriss wrote:
Sake Blok wrote:
On 4 mei 2011, at 22:11, Jeff Morriss wrote:
Max Dmitrichenko wrote:
Hi!
I'm continue to write dissector for an encrypted protocol. Everything
works fine until I receive an out-of-order TCP segment, i.e. previous
was lost.
Since I'm trying to decrypt it, I fail with it and break the whole
decryption context. Is there any way to:
1) Detect that this packet is out of order in given conversation?
2) Ask the TCP dissector to feed this packet later again when all
previous segments will be retransmitted?
I would think desegment_tcp() should be able to handle this by not calling your dissector for an out-of-order 
segment: it should be able to only call your dissector once it has a completely reassembled (desegmented) PDU.  
Looking through the code, it's not immediately obvious to me what the problem is.
One case that can cause a problem is when the first segment of a PDU is received out-of-order. Or did your recent 
work also handle this exception, Jeff?

Yep, that's the case rev 36304 fixed.

Unfortunately not, I just constructed a file in which the first segment of the HTTP-response PDU comes after the second 
segment of that PDU. And dissection still fails as there is not yet a segment in tcpd->fwd->multisegment_pdus for that 
seq. 

This could be fixed by adding a segment that comes after a gap to the multisegment list by default, but that would 
break dissection of single segment PDU's that come out-of-order. I don't know of a good way to solve this dilemma.

Cheers,


Sake

___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <wireshark-dev () wireshark org>
Archives:    http://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev
             mailto:wireshark-dev-request () wireshark org?subject=unsubscribe


Current thread: