Re: Buffer overflow in Python code

From: Ryan Permeh (ryanat_private)
Date: Mon Nov 26 2001 - 10:45:20 PST

  • Next message: Magniat_private: "Malicious use of grc.com"

    yes and no,  we released a sort of similar bug in asp, where feilds were
    overflowable.  using unicode, we were able to upload a asp script and cause
    an overflow that executed in system privs (unicode was not system).  I know
    there are python ports to win32, so this could apply there too.  you need to
    be able to get a script there in the first place, but then you may be able
    to do more, perhaps at a higher context than what you need to upload a
    script.  This may also be applicible to any type of embedded python system,
    perhaps used by a suid program.  I know perl is embeddable, and i blieve
    that python is as well.
    Signed,
    Ryan Permeh
    eEye Digital Security Team
    http://www.eEye.com/Retina -Network Security Scanner
    http://www.eEye.com/Iris -Network Traffic Analyzer
    http://www.eEye.com/SecureIIS -Stop Known and Unknown IIS Vulnerabilities
    
    ----- Original Message -----
    From: "Chris Ess" <azarinat_private>
    To: "Giorgio" <denebat_private>
    Cc: <vuln-devat_private>
    Sent: Sunday, November 25, 2001 10:05 AM
    Subject: Re: Buffer overflow in Python code
    
    
    > > I've found buffer overflow in Python 2.1.1 source code. (Maybe
    > > there're many others) The buffer overflow is in the file traceback.c
    > > in the directory Python of the Python source code.
    > >
    > > Simply there's a sprintf done in this way:
    > > sprintf(linebuf,FMT,filename,lineno,name) What cause the overflow is
    > > the name parameter which could be > 1000 (linebuf size) Alex Martelli
    > > <aleaxat_private> has submitted the bug on sourceforge as 485175, and
    > > produced the follow script to demostrate the overflow:
    >
    > Using the supplied script, I did achieve a segfault during the traceback
    > with Python 2.1.  However, I'm hardpressed to figure out how one would
    > exploit this...  After all, the Python binary is rarely SUID or SGID.  (I
    > know it's not on my system.)
    >
    > Is this a bug in the code?  Yes.
    >
    > Is this a security concern?  Right now, I'm inclined to say 'no'.  However
    > if it is, I would appreciate being told why.
    >
    > Sincerely,
    > Chris Ess
    > System Administrator / CDTT (Certified Duct Tape Technician)
    >
    >
    >
    



    This archive was generated by hypermail 2b30 : Mon Nov 26 2001 - 11:00:43 PST