Skip to contentSkip to conversation

Crypto operation

dm-crypt octane indice 3 participants 4 messages Nov 5, 2010 – Nov 7, 2010 2 messages not in the archive

lore
Share & tools
lore
#0 lore raw

Hello,

In order to better understand how cryptsetup works, I'm asking if there is a 'big picture' or a debug mode explaining how things are done. (I'm asking about cryptsetup, not LUKS).

e.g. An explanation as. Password is taken from command line.
A read is made through /dev/mapper/crypted_disk
The first block is read. The AES-128-cbc operation is computed with options like, iv= K= or salt=...
When we write a block, then which things are done, and so on.

Is there a tech documentation explaining that? The goal for me would to use openssl in order to make a decryption of a disk ciphered by cryptsetup and vice-versa.

In a second time, I would think that it works the salme way with Luks, excepts that the password is read from the luks header nad crypto operation begins beyond the luks header.

Thanks

-----------------------------------

Envoyé avec Inmano, ma messagerie renversante et gratuite : http://www.inmano.com

#1 lore raw
Quoted 3 lines · octane indice

In order to better understand how cryptsetup works, I'm asking if there is a 'big picture' or a debug mode explaining how things are done. (I'm asking about cryptsetup, not LUKS).

Then you are not asking about cryptsetup, but just about how dm-crypt works:)

Cryptsetup just configures dm-crypt, there is no userspace crypto related (except LUKS operation or passphrase hashing in plain mode
- output of this is just volume key directly used in dm-crypt)

Start with kernel documentation
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blob_plain;f=Documentation/device-mapper/dm-crypt.txt;hb=HEAD

you can get active mapped target (after cryptsetup configures it) with dmsetup table --showkeys

you can configure it using dmsetup too
(dmsetup is just wrapper around dm-ioctl calls)

Quoted 5 lines

e.g. An explanation as. Password is taken from command line.
A read is made through /dev/mapper/crypted_disk
The first block is read. The AES-128-cbc operation is computed with options like, iv= K= or salt=...
When we write a block, then which things are done, and so on.

that's not job for cryptsetup, this is how dm-crypt works

Example: in plain mode:
- cryptsetup reads passphrase, hashes it using specified hash alg and use it directly as volume key

- in luks mode it decrypts keyslot and read volume key from there

The enryption in dm-crypt is common mode for storage FDE, each sector (512B) is encrypted separately with given cipher and mode) and initial vector.

E.g. aes-cbc-plain : used cipher AES, cbc mode (so the 512B sector is divided to 16bytes cipher blocks for CBC) and with initial vector "plain" what here means just sector number.

Quoted 3 lines

Is there a tech documentation explaining that? The goal for me would to use openssl in order to make a decryption of a disk ciphered by cryptsetup and vice-versa.

That makes no sense to me (if you are using Linux), except some exercise with openssl.
What is the real goal here? Can you please explain it?

Quoted 3 lines

In a second time, I would think that it works the salme way with Luks, excepts that the password is read from the luks header nad crypto operation begins beyond the luks header.

passphrase in LUKS mode is neither derived nor related to volume key used in encryption. passphrase just unlocks the slot where the key is stored.

Milan

A message that is not in the archive (sent off-list or not yet received): [email protected]

A message that is not in the archive (sent off-list or not yet received): [email protected]

#4 lore raw

En réponse à Milan Broz <[email protected]> :

Quoted 4 lines
Quoted 1 line

Ok, what is this hash alg?

cryptsetup --help (with cryptsetup 1.1.3) shows defaults:

Ok, thanks

Quoted 3 lines

anyway, read this
http://www.markus-gattol.name/ws/dm-crypt_luks.html

yes, interesting

Quoted 2 lines

you did so many mistakes here.... that you provoke me to try it myself:-)

:-)

Quoted 1 line
 # openssl aes-128-cbc -nosalt -d -in dm.img -K

6469726563746b657900000000000000 -out dmc.img -iv 0 -p

Thank you very much, it will help me a lot.

Quoted 1 line

------------------- Fin du message d'origine ---------------------

Envoyé avec Inmano, ma messagerie renversante et gratuite : http://www.inmano.com

#5 lore raw
In reply to octane indice
Quoted 5 lines · octane indice

Hello,

In order to better understand how cryptsetup works, I'm asking if there is a 'big picture' or a debug mode explaining how things are done. (I'm asking about cryptsetup, not LUKS).

As MIlan already wrote, it is basically just the application of two crypto algorithms, both of which are not in cryptsetup, but in the kernel. As they are used in the context of the device mapper, they need to be in the kernel, but if you do direct access via an userspace application, they can be anywhere.

For your experiments, I recomend that you use a small loop-file, mount it via dm-crypt and put some plaintext in it, e.g. like this:

head -c 1M /dev/zero > cfile # create empty file
losetup /dev/loop0 cfile
cryptsetup create c1 /dev/loop0
man bash > /dev/mapper/c1
cryptsetup remoce c1
losetup -d /dev/loop0

You can then trial decrypt cfile. You can use --cipher, and --hashto set something simpler than the curent defaults of aes-cbc-essiv:sha256, ripemd160, but that should not make too much of a difference.

You can read about cipher modes here, if you want to do the mode yourself:
http://en.wikipedia.org/wiki/Block_cipher_modes_of_operation

ESSIV is described here:
http://en.wikipedia.org/wiki/ESSIV#ESSIV
with the hash that listed after the ':' in the cipher spec.

Arno

Signature · 7 lines
Arno Wagner, Dr. sc. techn., Dipl. Inform., CISSP -- Email: [email protected] 
GnuPG:  ID: 1E25338F  FP: 0C30 5782 9D93 F785 E79C  0296 797F 6B50 1E25 338F
----
Cuddly UI's are the manifestation of wishful thinking. -- Dylan Evans

If it's in the news, don't worry about it.  The very definition of 
"news" is "something that hardly ever happens." -- Bruce Schneier