Re: WinNT Widespread Teardrop Exploit

From: Russ (Russ.Cooperat_private)
Date: Wed Mar 04 1998 - 00:43:25 PST

  • Next message: Aleph One: "Update on wide-spread NewTear Denial of Service attacks"

    So far, on sites where caps were available (or tcpdumps) all replays of
    said caps have failed to crash machines patched against Teardrop2.
    
    Since its impossible to be certain, in such a short period of time (the
    attacks began on Sunday night EST and have continued through to the time
    of posting), that all attacks are the same (or significantly similar)
    there is hesitancy to say this is definitely Teardrop2.
    
    Some sites have reported DNS attacks, the sites I've talked to that saw
    attacking packets labeled DNS all indicated that those packets were, in
    fact, invalid DNS packets. Instead, it appears that fragmented UDP 53
    packets are being used to form the exploit and trigger the kernel crash
    on NT and Win95 boxes that have not been patched.
    
    At least one site reported that Linux kernels prior to 2.0.32 that have
    not been patched will freeze, this is consistent with Teardrop2.
    
    Win98 beta 3 machines seem to be unaffected, they include the Teardrop2
    fixes.
    
    I've had two confirmations, in addition to Microsoft, from very large
    orgs that machines patched with the Teardrop2 patch from January
    (identified in Dale's message) withstood attacks.
    
    Some valuable data points (again, at the time of posting);
    
    - Virtually all of the larger attacks seem to be originating from
    199.0.154.13, however this address is spoofed.
    
    - Many of the attacks seem to originate from a source port of 4000 and
    go after random ports. ICQ is on port 4000 but is, currently, not
    suspect.
    
    - The majority of other reports indicate source and destination ports
    53.
    
    - Everyone is seeing fragmented UDP packets with a 32 byte offset.
    Assembled size seems to vary, although this could just be a result of
    the analysis methods.
    
    The focus on .gov and .edu sites seems consistent with Aleph's story
    pointers.
    
    You can have a look at my NTBugTraq archives for today to follow my
    reports on the discoveries (as well as my silly theories).
    
    http://listserv.ntbugtraq.com/SCRIPTS/WA-NTBT.EXE?S2=ntbugtraq&q=&s=&f=&
    a=3+mar+98&b=4+mar+98
    
    Cheers,
    Russ
    http://www.ntbugtraq.com
    



    This archive was generated by hypermail 2b30 : Fri Apr 13 2001 - 13:43:53 PDT