From: =?iso-8859-1?Q?J=F6rn?= Engel Subject: Re: partially encrypted filesystem Date: Thu, 4 Dec 2003 19:40:34 +0100 Sender: linux-kernel-owner@vger.kernel.org Message-ID: <20031204184033.GA14029@wohnheim.fh-wedel.de> References: <1070485676.4855.16.camel@nucleon> <20031203214443.GA23693@wohnheim.fh-wedel.de> <20031204141725.GC7890@wohnheim.fh-wedel.de> <20031204172653.GA12516@wohnheim.fh-wedel.de> <3FCF7AD5.4050501@lougher.demon.co.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Kallol Biswas , linux-kernel@vger.kernel.org, "linux-fsdevel@vger.kernel.org" Return-path: To: Phillip Lougher Content-Disposition: inline In-Reply-To: <3FCF7AD5.4050501@lougher.demon.co.uk> List-Id: linux-fsdevel.vger.kernel.org On Thu, 4 December 2003 18:20:05 +0000, Phillip Lougher wrote: > J=F6rn Engel wrote: > > > >So - as sick as it sounds - jffs2 may actually be the fs of choice > >when doing encryption, even though working on a hard drive and not > >flash. Cool. :) > > >=20 > Considering that Jffs2 is the only writeable compressed filesystem, y= es.=20 > What should be borne in mind is compressed filesystems never expect=20 > the data after compression to be bigger than the original data. In t= he=20 > case where the compressed data is bigger, the original data is used=20 > instead, which is hardy ideal for an encrypted filesystem, and so mor= e=20 > than a direct substitution of compression function for encrypt functi= on=20 > is needed - this is of course only relevant if the encryption algorit= hm=20 > used could return more data... Correct. But this requirement can easily be weakened a enough for encryption to work. A couple ALIGN(..., encrypt_blocksize) at two or three places should do the trick. > >Depends on how much security you really care about. If you really > >don't mind the pain involved, some metadata should explicitly *not* = be > >encrypted, to avoid known plaintext attacks. To a serious attacker, > >this could be a death stroke for ext[23] over cryptoloop, actually. >=20 > You're assuming the metadata (inodes, indexes and directory entries),= =20 > are encrypted with the same key, and therefore decrypting the directo= ry=20 > data using plaintext attacks will give the attacker the key to the=20 > entire metadata? There is nothing preventing the directory data bein= g=20 > encrypted separately with a different key, and therefore a plaintext=20 > attack would get nothing more than the directory information. True, although that barely makes a difference. In either case you have to seperate known-plaintext data from the rest and either not encrypt it or use a different key. The hard part is seperating the data. > As you say, you highlight a drawback with cryptoloop and cloop, becau= se=20 > they cannot distinquish between different types of data. This sort o= f=20 > thing should always be done at the fs level rather than the block lev= el... If it is done, yes. "fsck it, who cares" may also be a valid design. > >In real life, though, the humans are usually the weakest link, so th= is > >doesn't matter anyway. >=20 > Hmmm, why not give give up completely then? Because it is fun doing so or someone pays for it. And - more to the point - the humans should be the weakest link, that is bad enough already. No reason to make/leave things worse. J=F6rn --=20 To recognize individual spam features you have to try to get into the mind of the spammer, and frankly I want to spend as little time inside the minds of spammers as possible. -- Paul Graham