Re: Ircii-epic: Irc: another funny stuff. In some irc clients dcc

From: Ben Winslow (rainat_private)
Date: Sun Dec 20 1998 - 07:35:43 PST

  • Next message: James R Grinter: "Re: Verifying file data integrity using L6"

    ---------- Forwarded message ----------
    Received: from BlackHole.RainNet.Org (rainat_private [192.168.1.3])
            by Portal.RainNet.Org (8.8.8/8.8.8/Debian/GNU) with ESMTP id KAA26632
            for <rainat_private>; Sun, 20 Dec 1998 10:31:10 -0500
    Received: from listopher.concentric.net (listopher.concentric.net
        [206.173.119.117])
            by BlackHole.RainNet.Org (8.8.5/8.8.5) with ESMTP id KAA13517
            for <rainat_private>; Sun, 20 Dec 1998 10:31:23 -0500
    Received: (from majordom@localhost)
            by listopher.concentric.net (8.8.3/8.8.5)
            id KAA21767; Sun, 20 Dec 1998 10:06:15 -0500 (EST)
    Message-ID: <199812201506.JAA27379at_private>
    To: ircii-epicat_private
    Subject: Re: Ircii-epic: Irc: another funny stuff. In some irc clients dcc may
        be hijacked.
    In-Reply-To: Your message of "19 Dec 1998 22:17:00 +0200."
        <77AMlEdphjBat_private>
    Date: Sun, 20 Dec 1998 09:06:07 -0600
    From: Jeremy Nelson <jnelsonat_private>
    Sender: owner-ircii-epicat_private
    Precedence: bulk
    
    >I just found a funny bug playing with some irc-client. DCC-chat may be
    >hijacked...
    
    This is not a bug in the client.  It is a function of the operating system.
    For example, this ``bug'' is not present in OpenBSD because it hands out
    ports randomly
    
    >The trouble comes while clients bind port to accept or request a dcc
    >CHAT/SEND/ or RECEIVE.  Being this a simple TCP connection without any ip
    >control.. the way to exploit is trivial.
    
    This is preposterous.  The client informs you of the remote IP address
    connecting.  Any half-aware user checks the IP address to make sure
    that it is reasonable.
    
    >Here we go:
    >
    >B , the hi-jacker wants to have fun with A. So he first creates
    >a dcc connection with A, getting the port binded.
    >
    >Now A is under attack since next ports used to create connections will
    >be quite consecutive to the first one. BitchX and IRCepic seem to be
    >affected with this matter. ( other clients???)
    >
    >Now A tries to /dcc chat C, but this is just a bit lagged. ( C maybe a
    >bot? ) B , using the following source, is going to assume the identity of C
    >except for his host. :-)
    
    Folks, this is completely preposterous.  This "exploit program" is nothing
    more than a limited-range port scanner.  What this "exploit" boils down
    to is:
    
            "If you establish a DCC connection with me, then if I port-scan
             you later between when you offer a DCC and when it is received,
             I will be able to connect to your DCC offer."
    
    Well, duh.  You could just turn this into a full-blown scanner and scan all
    day for DCC connections if thats what you wanted to accomplish, and even
    such a scanner as that would work on OpenBSD, where ports are handed out
    randomly.
    
    Folks, this is not a bug, except to the extent that you completely ignore
    the IP address on your established DCC transactions.  If its not the right
    IP, close it and try again.  And email the abuse contact of the offending
    ISP about how their users are port scanning you.
    
    Sheesh.
    Jeremy
    



    This archive was generated by hypermail 2b30 : Fri Apr 13 2001 - 14:25:20 PDT