On Tue, Jun 22, 2004 at 09:22:06PM +0300, Jari Ruusu wrote: > > loop-aes is nice and it does support the usual algorithms via an > > external patch. > > Contrary to lies that some people keep spreading, loop-AES has never > required kernel patches. Well, i was talking about the loop-aes ciphers-patch (http://loop-aes.sourceforge.net/ciphers/ to be clear). But maybe that has changed, so that patch has been included in loop-aes? Or am i missing something? > > The big drawback is that mount, umount, losetup, swapon > > and swapoff has to be patched and recompiled and loop-aes and cryptoloop > > doesnt work with eachother. > > Loop-AES can use cryptoloop's broken and backdoored on-disk formats just > fine when told to do that using mount options. Cryptoloop just cannot be > used in more secure mode that loop-AES and ciphers packages provide. Yes, so why use loop-aes with unpatched mount, umount, etc if there are no benefits? Hence, to use loop-aes you'll need to patch and recompile or you could just stick with cryptoloop. I could have been more clear when it came to the compatibility issue, but what i meant was that a loop-aes encrypted loop cant be decrypted with cryptoloop or vice versa. > Optimized dictionary attack: > http://marc.theaimsgroup.com/?l=linux-kernel&m=107419912024246&w=2 > > Watermark attack: > http://marc.theaimsgroup.com/?l=linux-kernel&m=107719798631935&w=2 Indeed serious but, and i really love a good conspiracy, isnt calling it backdoors a bit harse? /Thomas -- == thomas@xxxxxxxxxxxxxxxxxxxx | thomas@xxxxxxxxxxxx == Encrypted e-mails preferred | GPG KeyID: 114AA85C --
Attachment:
signature.asc
Description: Digital signature
