From: Matt Mahoney on
http://www.storagenewsletter.com/news/security/passware-kit-forensic

Any idea how this works?
From: Xavier Roche on
Matt Mahoney a écrit :
> http://www.storagenewsletter.com/news/security/passware-kit-forensic
> Any idea how this works?

According to the site's description, it works on a live PC with a
truecrypt partition mounted, when the PC is "locked".

The technic claimed may exploit a firefiwre flaw (see for example
http://www.hermann-uwe.de/blog/physical-memory-attacks-via-firewire-dma-part-1-overview-and-mitigation),
and is supposed to be able to get the keys in the dumped memory.

However, the "password recovery algorithms" for encrypted files technic
is a bit vague and lacks description. Dictionnary attack, maybe ?
From: Anne Onime on
"Xavier Roche" <xroche(a)free.fr.NOSPAM.invalid> wrote in message
news:hpno6s$hk$2(a)news.httrack.net...
> Matt Mahoney a �crit :
>> http://www.storagenewsletter.com/news/security/passware-kit-forensic
>> Any idea how this works?
>
> According to the site's description, it works on a live PC with a
> truecrypt partition mounted, when the PC is "locked".
>
> The technic claimed may exploit a firefiwre flaw (see for example
> http://www.hermann-uwe.de/blog/physical-memory-attacks-via-firewire-dma-part-1-overview-and-mitigation),
> and is supposed to be able to get the keys in the dumped memory.
>
> However, the "password recovery algorithms" for encrypted files technic
> is a bit vague and lacks description. Dictionnary attack, maybe ?

It may be useful in a few isolated cases, but for 99.9% of all
Truecrypt volumes (on powered down hard disks or flash disks) this will
be worthless unless the user encrypted with a weak password. In other
words: I'm not worried, I always log off when I leave my PC for any
length of time and I never leave it in standby mode. I currently don't
use Full Disk Encryption, but I keep thinking about it more and more
since it prevents someone from installing a keylogger on your machine.



From: nemo_outis on
Matt Mahoney <matmahoney(a)yahoo.com> wrote in news:323624e9-3a21-4a5d-
9601-1c31cd01d178(a)i25g2000yqm.googlegroups.com:

> http://www.storagenewsletter.com/news/security/passware-kit-forensic
>
> Any idea how this works?

Yep, the reference cited above explains how (well, drops a strong hint)
in its third paragraph:

"In response to customer requests, especially from law enforcement
organizations, Passware has enhanced Passware Kit Forensic to allow for
memory acquisition of a seized computer over FireWire port, even if the
computer is locked. When a target computer is seized and turned on with
the encryption disk accessible, the software scans its memory image and
extracts the encryption keys, so law enforcement personnel can access the
stored data."

IOW it's the straightforward Firewire DMA attack (metlstorm gets the
credit) that I wrote about here years ago (Feb 2008 and earlier) -
probably "tarted up" with a better interface. Moreover, this is NOT a
true attack against Truecrypt (or any similar program) since those
programs are designed to protect data "at rest" (i.e., unmounted) and NOT
while the key is in memory on a running machine. (It's really just an
attack on the rather weak keyboard lockout mechanism, not on Truecrypt).

Memory is then harvested and the key is found therein (some programming
required but nothing tricky).

Regards,

PS Most current attacks against encryption DO NOT try to break the
algorithm (e.g., AES256) or brute force the key; instead they attempt to
crack the password rather than the key (Most users use passwords much
weaker than the endelying algorithm/key).

Truecrypt (and several similar programs) use a method to hamper direct
attacks on the password - they make the relationaship between the
password and key computationally expensive. Which is to say the
(potential) relationship between a password and key is not 1:1 but
greater (enormously greater!). Making password cracking computationally
expensive typically involves salts and iterated hashing. It effectively
makes precomputation (rainbow tables, etc.) unfeasible and leaves no
better method than the near-brute-force method of trying each guessed
password to see if it yields a workable key. This method is pretty slow
and will only work if the user has chosen a short or weak password (e.g.,
a word or some simple variant). However, as said above, enough users do
select such weak passwords that even this slow dictionary attack is often
successful. But if you pick a strong password (or better, passphrase)
you're immune to it.

PPS To avoid the Firewire DMA attack you should disconnect all Firewire
ports (paranoids will put a blob of epoxy over them on the motherboard).
A hardware disconnect is far more reliable than a software one (e.g.,
disabling Firewire in the OS) but even the software disconnect will
thwart all but the most serious adversaries.

There are also some possible USB attacks that use similar methods to the
Firewire attack. Fortunately the USB attacks are much less powerful and
so I'll leave it to your discretion whether to disable USB on your system
(USB is so convenient as to be virtually a necessity). A software USB
disabler (there are several such programs) may be a satisfactory
compromise. FWIW I disable Firewire (in hardware) but not USB.


From: Maaartin on
On Apr 9, 11:53 pm, "nemo_outis" <a...(a)xyz.com> wrote:
> There are also some possible USB attacks that use similar methods to the
> Firewire attack.  Fortunately the USB attacks are much less powerful and
> so I'll leave it to your discretion whether to disable USB on your system
> (USB is so convenient as to be virtually a necessity).  A software USB
> disabler (there are several such programs) may be a satisfactory
> compromise.  FWIW I disable Firewire (in hardware) but not USB.

What about SATA? It's capable of DMA as well and there're often
external connectors for it.