Re: New vulnerability in IIS4.0/5.0

From: Paul McGovern (islesat_private)
Date: Thu Sep 20 2001 - 20:58:36 PDT

  • Next message: Seth Arnold: "Re: Question about Local vulnerability in libutil derived with FreeBSD."

    On Wed, 19 Sep 2001, Dave Ahmad wrote:
    
    | This seems to be just be another way to exploit the double decode
    | vulnerability (Bugtraq ID 2708).  There is a possibility that it may be a
    | new issue due to the use of '%u' method of encoding.  It does not look
    | that way to us.
    |
    | <snip>
    |
    | Has anyone managed to exploit a patched system?
    
    Unfortunately, I have. I noticed a few weeks back that our network at
    work was periodically getting extremely slow, and after a bit of
    investigation utilizing tcpdump, it turned out our NT4 webserver (running
    IIS4 with all up-to-date security patches) was being used to pingflood
    various hosts with the exact exploit mentioned in the advisory which
    started this thread. The IIS logs showed what translated into the
    following:
    
    http://target/scripts/..%u005c..%u005cwinnt/system32/cmd.exe?/c+ping+ ...
    etc
    
    Baffled, I double-checked to make sure the decode vuln. patch had been
    installed, and it was indeed there. After trying to reapply the patch, I
    figured IIS just wasn't taking the patch and did a stopgap fix using some
    file renaming and guest access permission-removal tricks. The machine in
    question is being upgraded to Win2k server very soon anyway, so the
    stopgap was good enough for the past few weeks. I suppose my assumption
    that there was a problem with our IIS4 installation (causing the hotfix
    not to work) may have been incorrect after reading this advisory. Your
    mileage may vary :)
    
    --
    Paul McGovern
    http://isles.krad.org
    



    This archive was generated by hypermail 2b30 : Fri Sep 21 2001 - 12:11:40 PDT