[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <53D2BDF3.5060909@cs.unm.edu>
Date: Fri, 25 Jul 2014 14:28:35 -0600
From: Jeffrey Knockel <jeffk@...unm.edu>
To: Eric Dumazet <eric.dumazet@...il.com>
CC: Linus Torvalds <torvalds@...ux-foundation.org>,
David Miller <davem@...emloft.net>,
netdev <netdev@...r.kernel.org>,
"Jedidiah R. Crandall" <crandall@...unm.edu>,
Willy Tarreau <w@....eu>,
"security@...nel.org" <security@...nel.org>
Subject: Re: [PATCH net] ip: make IP identifiers less predictable
On 07/25/2014 12:09 PM, Eric Dumazet wrote:
> What do you mean by "an attacker who controls a large number of
> addresses" ?
In general, I mean any attacker who can read the packets sent to a large
number of different Internet addresses. Even an attacker who has one
address but can cycle through different assignments may be a problem.
> The hash(daddr) -> slot function is not known, as we use a Jenkin hash
> with a secret ( ip_idents_hashrnd & ip6_idents_hashrnd )
That's true, but the secret never changes, right? I may not be able to
identify the slot number that any address is hashed to, but I can
identify when some victim address hashes to the same slot as one of my
addresses whose packets I can read. For instance, if I in short succession
1. Probe value of the IP id counter for each of my addresses
2. Spoof a large number of (e.g.) echo requests from victim address (or
something else to the distribution that I can measure)
3. Again probe value of the IP id counter for each of my addresses
Then I can tell which of my addresses hash to the same slot as the
victim address by whose value of the IP id counter has jumped as a
result of the linux machine sending echo replies to the victim.
Jeff
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists