* markus reichelt <ml@xxxxxxxxxxxxx> [2007-05-28 15:01:05 +0200]: > Bad choice of words in the manual, imho. Don't worry about this. Agreed. > in that example they use kernel 2.6.8 (hint hint) > " ... not a big problem." = dont worry about this. > .... But in case an attacker has access to your > ciphertext you already got a bigger problem. > > So, what are the current problems of each implementation (if any)? > by now you are fine with any of the stooges. Just don't use blowfish > cipher, it is not recommended for large amounts of data. Agreed, however: Here I must point out that although I prefer loop-aes over any other for full disk encryption, initial implementation requires a bit more in terms of discipline for successful application. If it's done properly, algorithm and key structure are never revealed to the adversary, making an attack much more difficult to mount. The downside is the extra work involved in performing the setup and maintenance of the system, so a weakness is introduced in the form of a human engineering attack. Cryptsetup-luks is easier to implement and maintain, but it does reveal the algorithm to the adversary at the outset to anyone who examines the disk with a luks dump command. The human risk factor is greatly reduced, however, in the cryptsetup implementation - it's harder to screw up. > well, apart from the usual mantra about backups, all my updates have > gone smoothly so far (I'm solely using loop-aes). Same here. Regards, Unweitze - Linux-crypto: cryptography in and on the Linux system Archive: http://mail.nl.linux.org/linux-crypto/
