Re: ipop3d (x2) / pine (x2)

From: Oliver Xymoron (oxymoronat_private)
Date: Sun Apr 11 1999 - 09:00:09 PDT

  • Next message: GvS: "Re: ipop3d (x2) / pine (x2) / ..."

    On Thu, 8 Apr 1999, Mark Crispin wrote:
    
    > Now, we'll talk about the 20% that is fact.  Yes, it is possible to write
    > a negative process ID in the lock file.  This requires that the attacker
    > have shell access; it can't be mounted remotely.  It also requires that
    > the attacker have a program running at the time that the victim opens his
    > mail file.
    
    Attackers sometimes have shell access. Use your imagination. Attackers can
    sometimes conceal attempts to exploit races long enough for it to work.
    Use your imagination. I for one am not always running top.
    
    > Not only is the program running, but it has the lock file open and
    > locked.  In other words, it's incredibly easy to catch; particularly
    > if you have lsof. Nor can there be any question of intent when it
    > comes to prosecution.  Only an extremely stupid individual would try
    > it.
    
    Imagine this scenario. You have a thousand users. One of them is stupid
    enough to be duped out of their password. Odds? High. Attacker gets in,
    connecting from a machine that they've already 'owned.' Scenario two. You
    have a thousand users. One of them uses telnet/pop/imap/etc to connect
    from a machine through a route that's being snooped from another owned
    machine. Perhaps they've tunneled through a bunch of poorly maintained
    UNIX boxes in university dorms and are now on your LAN. Odds? High.
    
    Risk to attacker of shell experimentation in the early morning hours?
    Negligible.
    
    Exercise for the reader: invent a scenario whereby an attacker engineers
    their way into root access by having access to unpriviledged accounts. I
    can think of five off the top of my head and it's only 11am on a Sunday.
    
    In short, own your bugs. There are no trusted users, especially on a
    networked machine. Users have to be protected from other users. Period.
    That's the security model. If Pine lets an attacker make progress
    through a system, that's a security hole _in Pine_. Deal with it.
    
    The past few posts to Bugtraq I've seen from the Pine group have rather
    worried me. Consistently taking a posture of "we don't think it's a
    problem" or "it's not our problem" does little to inspire community
    confidence. Your code will have bugs or bad interactions with other
    programs. They will be found, especially since Pine is popular. Some of
    them will be posted publically, especially if you don't appear to take a
    proactive stance. Be gracious about it.
    
    --
     "Love the dolphins," she advised him. "Write by W.A.S.T.E.."
    



    This archive was generated by hypermail 2b30 : Fri Apr 13 2001 - 14:42:03 PDT