[dm-crypt] FAQ Montly pointer 12/2012

[dm-crypt] FAQ Montly pointer 12/2012

[Arno Wagner]

Updates: Added a passphrase strength estimate based on
http://it.slashdot.org/story/12/12/05/0623215/new-25-gpu-monster-devours-strong-passwords-in-minutes

Executive summary: It may be time to increase passphrase length.

-----

The FAQ can be found at
- http://code.google.com/p/cryptsetup/wiki/FrequentlyAskedQuestions
and 
- in the sources as FAQ in the top-level directory.

-- 
Arno Wagner,     Dr. sc. techn., Dipl. Inform.,    Email: arno@wagner.name
GnuPG: ID: CB5D9718  FP: 12D6 C03B 1B30 33BB 13CF  B774 E35C 5FA1 CB5D 9718
----
One of the painful things about our time is that those who feel certainty
are stupid, and those with any imagination and understanding are filled
with doubt and indecision. -- Bertrand Russell

Re: [dm-crypt] FAQ Montly pointer 12/2012

[Javier Juan Martínez Cabezón]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 06/12/12 16:29, Arno Wagner wrote:
> Updates: Added a passphrase strength estimate based on 
> http://it.slashdot.org/story/12/12/05/0623215/new-25-gpu-monster-devours-strong-passwords-in-minutes
>
>  Executive summary: It may be time to increase passphrase length.
> 
> -----
> 
> The FAQ can be found at -
> http://code.google.com/p/cryptsetup/wiki/FrequentlyAskedQuestions 
> and - in the sources as FAQ in the top-level directory.
> 



This is applicable if the algoritm used to cipher is unknown as happen
in plain dmcrypt?


Arno, is really useful/secure to get the ciphering algoritm in the
LUKS header?

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)
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=RCwp
-----END PGP SIGNATURE-----

Re: [dm-crypt] FAQ Montly pointer 12/2012

[Arno Wagner]

On Fri, Dec 07, 2012 at 08:29:09PM +0100, Javier Juan Martínez Cabezón wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> On 06/12/12 16:29, Arno Wagner wrote:
> > Updates: Added a passphrase strength estimate based on 
> > http://it.slashdot.org/story/12/12/05/0623215/new-25-gpu-monster-devours-strong-passwords-in-minutes
> >
> >  Executive summary: It may be time to increase passphrase length.
> > 
> > -----
> > 
> > The FAQ can be found at -
> > http://code.google.com/p/cryptsetup/wiki/FrequentlyAskedQuestions 
> > and - in the sources as FAQ in the top-level directory.
> > 
> 
> 
> 
> This is applicable if the algoritm used to cipher is unknown as happen
> in plain dmcrypt?

It depends. In plain dm-crypt, determining whether you have
guessed the password is tricky anyways. Testing for different
ciphers would add some overhead and maybe give you something 
like 5-6 bits of security in addition.

> 
> Arno, is really useful/secure to get the ciphering algoritm in the
> LUKS header?

Yes, very much so. For LUKS, you do trial decryption of a keyslot,
and that is enough to tell you whether you have a good password.
My intuition is that the iterated hashing takes a lot longer than
the trial decryption, so trying different ciphers does not
add significant overhead. And there are not that many possible
ciphers anyways. Now, if you decide to not store the cipher,
then normal decryption would also need to try them and the
time would need to be taken into account in the maximum delay
a user is faced with. Better to increase the hash iteration 
count instead. 

Bottom line: An unknown hash or cipher algorithm does not add
much, unless it is homebrew. But in that case, it is still
accessible via the kernel and glaringly obvious. Besides, 
homebrew ciphers are often insecure.

Arno
-- 
Arno Wagner,     Dr. sc. techn., Dipl. Inform.,    Email: arno@wagner.name
GnuPG: ID: CB5D9718  FP: 12D6 C03B 1B30 33BB 13CF  B774 E35C 5FA1 CB5D 9718
----
One of the painful things about our time is that those who feel certainty
are stupid, and those with any imagination and understanding are filled
with doubt and indecision. -- Bertrand Russell