Re: partman-crypto
On Fri, Aug 05, 2005 at 09:05:52AM +0100, Martin Michlmayr wrote:
> Thanks for working on this. Just a quick question: what's the
> advantage of loop-AES over dm-crypt?
That's of course a matter of discussion. :-)
loop-AES has a very good track record with regards to introducing
measures against possible attacks. For example, the 2.x versions
introduced a multi-key mode to protect against a chosen plaintext
attack that had been published shortly before.
Due to this multi-key mode and a stronger method for deriving IV,
loop-AES is thought not to be vulnerable to a number of non-fatal
crypto attacks that cryptoloop and (last time I looked) dm-crypt
have no protection against. From what I read, the LUKS version of
dm-crypt has some improvements over plain dm-crypt in this regard.
There are also a couple of advantages from a usability POV:
Encryption keys are wrapped in GnuPG keyfiles and their passphrases
can be changed without having to re-encrypt the device[1]. Then
there is a portable userspace tool that can decrypt loop-AES volumes
on most POSIX systems without requiring kernel support.
> Also, without having looked at your code, do you think your package is
> flexible enough to allow for different encryption methods?
I'm not actually very familiar with dm-crypt, but I've had this
in mind when I started to work on it.
Much of the code should be identical for dm-crypt and loop-AES. In
principle the package is also prepared for using dm volumes and
handling their specific settings (keysize, hash function?, volume
name), only most of those things have not been implemented so far.
A first step could be to split the passphrase questions out of
loop-aes-keygen into a common file, fill in the currently empty
setup_dmcrypt() function in init.d/crypto and add a volume name
questions in active_partition/. (There might be complications from
partman assuming that everything in /dev/mapper/ is an LVM device,
but that was only from a quick look and I could well be wrong.)
Generally speaking, if someone is interested in completing the
dm-crypt and LUKS support in partman-crypto, I would be happy to
contribute and help get it working. I'm mainly lacking knowledge
and experience with either to do a good job with this myself.
Wesley I'm CCing you to ask for correction on what I wrote about
dm-crypt above :-P and if you'd perhaps be interested in joining
this work. We'd probably need at least a cryptsetup-udeb, if my
understanding of dm-crypt is correct.
cheers,
Max
--
[1] I suppose one could do similarily for dm-crypt, so this is not
really a genuine advantage.
Reply to: