Linux 2.4.27 SECURITY BUG - TCP Local and REMOTE(verified) Denial of Service Attack
Hey,
I'm not Alan, just trying to save him some typing :) The
issue being referenced is that BGP4 as you know uses TCP communication to check link status. If the TCP session is severed in any way, BGP assumes the link down, and drops the advertisements, or the table depending which side (at least in Cisco's implementation).
This basically leaves you dead in the water for a few seconds while the BGP session is re-established, and the advertisements are send out, and table rebuilt. It can also cause other routers on the net to see you as "flapping", and dampen your routes.. Which again leaves you dead in the water (at least from things behind them).
This is accomplished by guessing the correct TCP Sequence number, and sending RST packets to drop the TCP connection.
BGP does not actually check the layer 2 status of the connection to make sure the link is still UP before it assumes you have dropped. I believe this is the poor implementation he is referring to.
MD5 encryption was added to the sessions between routers to make hijacking the stream more difficult (if not next to impossible)
D.
Quoted 55 lines
-----Original Message-----
From: Dale Corse [mailto:[email protected]]
Sent: Sunday, September 12, 2004 1:36 PM
To: 'Dale'
Subject: FW: Linux 2.4.27 SECURITY BUG - TCP Local and REMOTE(verified) Denial of Service Attack-----Original Message-----
From: Toon van der Pas [mailto:[email protected]]
Sent: Sunday, September 12, 2004 1:24 PM
To: Alan Cox
Cc: Wolfpaw - Dale Corse; [email protected]; Linux Kernel Mailing List
Subject: Re: Linux 2.4.27 SECURITY BUG - TCP Local and REMOTE(verified) Denial of Service AttackQuoted 3 lines · Alan Cox
This is not a TCP flaw, its a combination of poor design by certain vendors, poor BGP implementation and a lack of
understanding of what
Quoted 6 lines
TCP does and does not do. See IPSec. TCP gets stuff from A to B in order and knowing to a resonable degree what arrived. TCP does not proide a security service.
(The core of this problem arises because certain people treat TCP connection down on the peering session as link down)
Alan, could you please elaborate on this last statement? I don't understand what you mean, and am very interested.
Thanks,
Toon.Signature · 13 lines
"Debugging is twice as hard as writing the code in the first place. Therefore, if you write the code as cleverly as possible, you are, by definition, not smart enough to debug it." - Brian W. Kernighan -------------------------------------------------------------- -------------- - This message has been scanned for Spam and Viruses by ClamAV and SpamAssassin -------------------------------------------------------------- -------------- -
Quoted 3 lines · Wolfpaw - Dale Corse
MD5 encryption was added to the sessions between routers to make hijacking the stream more difficult (if not next to impossible)
Correction : MD5 *signature* was added from the beginning since the problem was identified from start, but seeing that certain people did not implement it, others found interesting to turn this into a "generic TCP vulnerability" to get some credits, or perhaps to make them react positively.
Regards,
Willy