Versions of Cisco PIX firewalls do not validate the checksum of transiting TCP packets. Attackers may be able to use this problem to create a sustained denial-of-service under certain conditions.
Cisco PIX firewall systems are used to enforce site-specific network security policy. A problem related to a failure to validate the checksum information of TCP traffic by default may be used by remote, unauthenticated attackers to create a sustained denial-of-service against PIX-protected systems under certain conditions.
This condition may occur when TCP SYN packets with malformed TCP checksums and spoofed source addresses and port values are sent to systems behind affected PIX firewalls. Since the PIX does not validate the TCP checksum by default, it allows such packets through, creating an embryonic connection entry to track the connection attempt to the destination from the spoofed source address and port. The target of the attack would silently drop malformed TCP SYN packets without sending TCP RST packets back to the PIX to remove the embryonic connection entry. Legitimate attempts to connect to PIX-protected systems may then be blocked for up to two minutes per attack (assuming default embryonic connection timer settings).
Note this attack would not affect established TCP connections.
Exploit code has been made publicly available that may automate a sustained denial-of-service attack described above.
Valid TCP connection attempts originating from the spoofed source IP and source ports during a sustained attack may not be allowed through affected PIX firewalls.
Several workarounds are publicly described by Cisco:
Vendor| Status| Date Notified| Date Updated
Cisco Systems, Inc.| | -| 30 Nov 2005
If you are a vendor and your product is affected, let us know.
Group | Score | Vector
Base | N/A | N/A
Temporal | N/A | N/A
Environmental | N/A | N/A
This document was written by Jeff S Havrilla.