Some time ago I posted a rough script for layered encryption by chaining loop devices. In the meantime I have polished that script.
As one option, the script automatically generates (prompts for password) and embeds a gpg key in the encrypted loop device of the preceding layer The gpg key for the first layer is always an external key, which is the only key the user need keep externally.
It occurs to me that this might introduce a vulnerability. If the embedded key is ascii armoured, it always has the same string at the beginning "-----BEGIN PGP MESSAGE-----" etc. Or if it is binary, it has headers in a known format.
So, my question is this: could an attacker looking at a partition encrypted in this way (which appears random), but knowing that at so-many bytes from the beginning there might be a loop-aes encrypted gpg key which will either have the loop-aes encrypted acii plaintext string above or loop-aes encrypted gpg binary headers, use this knowledge to crack the first (top) layer of encryption?
-
Linux-crypto: cryptography in and on the Linux system
Archive: http://mail.nl.linux.org/linux-crypto/
