Jacek's Status Report #7 of 13

[Jacek Wielemborek <d33tah () gmail com>]

Hello,

This is my report for week 7 of "Nsock-based port scanning"
Google Summer of Code project.

Accomplishments:

* Split my API into nsock-specific and more general functions, which
will make it possible to implement raw scans in the future (also, it
makes the code look cleaner),

* Did some more cleanup in the code, such as fixing trivial bugs, adding
comments and fixing coding style,

* Made my regression testing program produce much more usable reports:
 -> Instead of putting lengthy logs directly to the e-mail body, I
implemented adding them as attachments,
 -> Removed testing of nmap-nseportscan as I'm not developing it anymore,
 -> Added code that keeps track of the number of file descriptors Nmap
has open at the given moment and plots a graph which compares these
values for both Nmap trunk and my branch. This lets me see how my
prototype's congestion control behaves,
 -> Unified the test cases by adding -n -Pn, added scanning of
scanme.nmap.org to measure performance on a real network as well,
 -> If SVN trunk and my prototype's port scanning results happen to be
incosistent, show the differences in "diff" format instead of printing
all the information. This only happens because of an Nmap bug with Linux
ephemeral ports right now,

* Tested --ip-options, experimented with --proxies (I'm currently
waiting for Henri's response on Nsock API),

* Fixed the "ephemeral"/"self-connect" bug that happens on Linux
systems. I used Daniel's code to solve that
(http://seclists.org/nmap-dev/2014/q1/136),

* Met with my mentor. Discussed - among other things - more regression
testing scenarios and an implementation of a more advanced program that
plots Nmap's congestion control performance,

* Modified my branch's debug output related to congestion control to
make it look like SVN trunk's. Wrote another plotting program which,
instead of looking at file descriptor count, parses Nmap's "*** TIMING
STATS ***" debug lines and extracts data from there. The program is not
included in the regression testing program yet,

* Added code that keeps track of per-host congestion control variables.
The data generated by Nmap's per-host ultra_timing_vals isn't used yet.

Priorities:

* Implement test cases for the new functionality I added last week and
perhaps also the functionality I plan to add soon,

* Use per-host congestion control structures to limit the maximum number
of probes sent per host,

* Keep debugging my congestion control code with the aim of making its
behavior consistent with SVN trunk's,

* Implement timing-related functionality I mentioned in last week's
priorities.

Yours,
Jacek Wielemborek

Attachment: signature.asc
Description: OpenPGP digital signature

_______________________________________________
Sent through the dev mailing list
http://nmap.org/mailman/listinfo/dev
Archived at http://seclists.org/nmap-dev/