[Snort-users] Session mixup by stream4

Sonali Gupta sonali.gupta at ...11827...
Tue Feb 8 01:32:02 EST 2005


I came across this post in the neohapsis archives, which discusses an
issue similar to what I seem to be facing.

It is at: 



The discussion talks about a session payload mixup in data captured by
snort. I am also facing this issue in some sessions that I get from
snort. I am using snort 2.0 and snort 2.3 ORC2, and found the issue in
both versions. Chris mentioned in the post that the issue has been
fixed in HEAD CVS. The discussion was about version 1.9.0. Since I am
using a newer version of snort, could you please tell me if the same
fix has been applied to the newer versions, or is there some update I
need to get to fix it.

In order to figure out the session mix up issue, I did the following

a)       Ran two snorts simultaneously Snort-A and Snort-B. The
Snort-A captures the traffic and dumps it in a tcpdump format, using
the –b option and Snort-B creates the sessions real time.

b)       Then I ran Snort-C (same executable as of Snort-B), which
created session from the tcpdump produced by Snort-A. I found that the
sessions created by Snort-C for the problematic ones were not
identical to the sessions created by Snort-B and I also noticed that
the sessions created by Snort-C has lesser number of session mixups.


I also noted that some sessions created by Snort-B, which appeared
incomplete (data from last one or two packets were missing) were fully
formed when created by Snort-C.


Is there any difference between the session reassemble real time and
session reassemble using –r option.


During the test, we had tweaked Snort-B a bit so that it dumps all the
sessions that are reassembled. We did this by hard coding gotevent = 1
and commenting the call to the preprocessor function in FlushStream
function of ssp_stream4.c.


//        gotevent = Preprocess(stream_pkt);  //Commented

          gotevent = 1;        



Sonali Gupta

More information about the Snort-users mailing list