[Bro-Dev] dns.bro

Robin Sommer robin at icir.org
Mon May 2 15:35:10 PDT 2011


As discussed, I'll start going through the scripts in the
policy-scripts-new branch, starting with dns.bro.

I really like the new dns.log, pretty neat!

Two general, DNS-independent, questions first:

    - Does a policy/foo.bro script always load all of
    policy/foo/*.bro? Would be nice if that was consistent, and
    perhaps it already is. :-)

    - We should include new connection$uid into pretty much all
   relevant logs.

dns/base.bro:

    - There are number of commented out "print" statements. Should we
      pass this into weird.bro?

    - The script activates the binpac analyzer. Do we want to remove
      "classic" C++ analyzer?

    - There's a TODO about the EDNS/TSIG. What's the problem?

    - The reply handlers check for "ans$answer_type == DNS_ANS", but
      there are also options dns_skip_all_auth/dns_skip_all_addl in
      bro.init? Can we get rid of one of the two ways (I'd say the
      latter)?

    - The reply handlers are all almost identical. How about
      refactoring that code into a function called by them all?

    - The comment in connection_state_remove() seems misleading: this
      is the only place that logs anything, right?

    Watislat_ctveusd or
  - om o te yps n ontsbr dn' semto Do we
      want to delete he?

dns/passive-replication.bro

    - Can you remind me what the passive replication is for? I thought
    I knew but not sure that's matching with the script. :-)

    - Regarding the TODO: should "recent_requests" be a table[string]
    of set[string]?  

Robin


--
Robin Sommer * Phone +1 (510) 722-6541 * robin at icir.org
ICSI/LBNL    * Fax   +1 (510) 666-2956 *   www.icir.org


More information about the bro-dev mailing list