Re: Wiping Swap Partitions

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Sun, May 11, 2003 at 06:28:29PM +0300, Jari Ruusu wrote:
> > >  /dev/hda666   none   swap   sw,loop=/dev/loop6,encryption=AES128   0   0
> > 
> > Is that work with kernelI (with losetup and mount) patch?
> > Any kernel Int cipher may be selected?
> 
> kerneli.org version does not have 'encrypted swap' enabled swapon+swapoff.

Yeah, I missed that. I handle the encryption only through losetup. The
keys are generated from /dev/urandom and are piped to losetup in my
scripts. Since I have 2 swap partitions I have "pri=1" in fstab options
to specify the priorities (equal, to distribute swap among both.) The
loop devices are listed in my fstab, not the underlying partitions; that
again is handled by losetup in my rc.local script.

So to answer the anonymous poster's query more fully: no, that type of
fstab entry will not work with the kerneli patch, but you can accomplish
the same thing using my scripts and "normal" fstab swap options. Thanks
again, Jari.

Sam, I think by now the original post in this thread has been beaten
pretty badly, but I had intended some days back to point you to a
discussion in comp.os.linux.security regarding "wiping" data from
magnetic media. It is conceivable that even attackers of modest means
could recover overwritten data, especially if overwritten with /dev/zero
putput.
  http://groups.google.com/groups?selm=m3u1kvdtar.fsf%40mika.informatik.uni-freiburg.de
My proposal had been to use badblocks to overwrite the data, and the
above-referenced post explains why even that approach, whilst better
than /dev/zero, would not provide much more security.

"FBI" is mentioned numerous times in the thread as the threat model, but
see the Gutmann paper also mentioned:
  http://www.cs.auckland.ac.nz/~pgut001/pubs/secure_del.html
Not "light reading", but just go down to section 2. Gutmann suggested in
1996 that a powerful data recovery system could be built very cheaply:
  "If commercially-available SPM's are considered too expensive, it is
  possible to build a reasonably capable SPM for about US$1400, using
  a PC as a controller"
Consider as well that he was saying this in 1996, with more valuable
dollars buying more expensive computing equipment. It might be much
lower in cost now -- certainly on much more powerful processors.

Suppose an attacker seizes physical control of your system before a
normal shutdown: your swap is readable. That's a low-tech, yet very
effective, attack. No matter what your threat model, I'd recommend that
you encrypt rather than wipe your swap devices.

    Rob - /dev/rob0
-
Linux-crypto:  cryptography in and on the Linux system
Archive:       http://mail.nl.linux.org/linux-crypto/


[Index of Archives]     [Kernel]     [Linux Crypto]     [Gnu Crypto]     [Gnu Classpath]     [Netfilter]     [Bugtraq]
  Powered by Linux