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

[Freedombox-discuss] PSN, ARM's Trust Zone and TPM

Hash: SHA1

On Thu, 28 Jun 2012, freebirds at hushmail.com wrote:

> --[PinePGP]--------------------------------------------------[begin]--
> Ben Mendis, you are missing my points. Regardless whether a
> product, such as software, ebook, video, etc. are purchased with
> DRM, the two UUIDs of TPM and the PSN are visible online to
> websites.

Because you haven't described _how_ they are "visible online".

> I already quoted that Intel's PSN is sent to Microsoft. When
> Windows computers start up, Microsoft automatically authenticates
> computes regarding whether they have genuine Microsoft. Microsoft
> antivirus and WMP does this too. Microsoft reads the PSN and TPM of
> computers to match the hardware with Microsoft' serial number.

Ok, so WGA uses it to authenticate machines when they connect to the
internet. Great. But FB isn't using Windows, so how is that relevant.

> There are articles that Microsoft's customers information is
> available to government. See
> http://newsworldwide.wordpress.com/2008/05/02/microsoft-discloses-
> government-backdoor-on-windows-operating-systems/

That's not a real news site, that's someone's blog. The author is making
a lot of wild speculation based on a few partial quotes.

> http://www.pcworld.com/article/190233/microsofts_spy_guide_what_you_
> need_to_know.html

I'll accept PCWorld, but the kind of information that it accuses
Microsoft of making available to law enforcement is the same kind of
information that law enforcement would typically be entitled to anyways
via subpoena. If you use an online service, such as Hotmail, then
someone else owns your emails. Your Fourth Amendment rights are forfeit.
Just ask the EFF. Again, this has noting to do with PSN/TPM.

> Microsoft and Skype's backdoor for government is at:
> http://memeburn.com/2011/07/microsoft-and-skype-set-to-allow-
> backdoor-eavesdropping/

- From the article, "I understand where Microsoft is coming from. They are
obliged, by law, to provide some sort of tracking tool for the
authorities who require these specific services. The US law, set by
CALEA (Communications Assistance for Law Enforcement Act), states that
all telecommunications operators must enable their hardware and software
for surveillance tracking."

Yes, lets all get mad at Microsoft for complying with the law. But, you
know, since we know that Microsoft own Skype maybe we should just stop
using it. Try one of the open source alternatives that don't come
bundled with a free backdoor program. Just a thought. Still not relevant

> TPM is not software dependent. "The TPM is bound to a single
> platform and is independent of all other platform components (such
> as processor, memory and operating system)."
> http://h20331.www2.hp.com/Hpsub/cache/292199-0-0-225-121.htm

That link gives me a 404.

> TPM is on by default. Users do not need to enable it.

On and doing... what, exactly? I get it, the TPM is an automous
subsystem. But what exactly does it _do_ on its own? And how? Can it
send and receive network traffic without the OS? Can it read data off
the hard drives without the OS? Can it access read and write to random
or even protected memory without the OS? If so, how does it do these
things. The technical specifics would need to be documented somewhere in
order for this platform to be useful even for the more legitimate of
the proposed usecases.

> TPM is not used only when users purchase a DRM product. Reread the
> list of ARM's TrustZone's users in my prior email.

I didn't say it was? I simply said that the DRM functionality is moot if
the reader application and the ebook document themselves are DRM-free.
The same goes for the rest of those usecases. Somewhere a developer
needs to make a conscious choice to enable the use of functionality
provided by the TPM/TrustZone chip.

> Website and malware use Javascript. Javascript can read UUIDs.
> Apple prohibits javascript in apps from reading UUIDs: "The uuid
> property returns the device?s unique identification id. NOTE: Apple
> no longer permits obtaining the uuid within applications. If you
> use this property in an app intended for Apple, it may get rejected
> or pulled from the store without notice at a later date. This
> property is still permitted for Android."
> http://www.appmobi.com/documentation/device.html

Ok, so for mobile devices the UUID can be read through JavaScript. Which
means that the browser and OS both need to support that. (JavaScript, on
it's own, can't talk directly to the BIOS, CPU or TPM. The JavaScript
interpreter would need to support doing that.) Also, if we're talking
about cell phones there are far more valuable serial numbers to capture,
such as the IMEI of the cellular radio.

Do desktop browsers also support this functionality in their JavaScript
implementations? Is there a JavaScript interpeter on the FB that
supports this functionality?

> "An examination of 101 popular smartphone "apps" ...
> showed that 56 transmitted the phone's unique device ID to
> other companies without users' awareness or consent. ...
> Five sent age,gender and other personal details to outsiders. ..."

Ok, sure. But how do a phone figure out the user's age and gender? I've
certainly never told my phone what my birthday or gender was. So how
does it know these details in order to report on me?

> Many apps written for smartphones are also written for tablets and
> PCs. They read the UUIDs of computers and sell this information.

UUIDs or IMEIs? I know there's a market for IMEIs, but I'd be surprised
if anyone was paing just for UUIDs.

> This week, Intel's processor was hacked again.
> http://thehackernews.com/2012/06/intel-cpu-vulnerability-can-
> provide.html

Again, this feels like misdirection. Yes, there was a flaw in an Intel
CPU, but the flaw was in the handling of the SYSRET instruction. That
instruction has absolutely nothing to do with PSN or TPM. So it doesn't
support your assertion that PSN and TPM are the problem.

> News articles on hacks do not give a step by step tutorial on how
> to to do. Hacking websites and forums may have tutorials. Visible
> PSN enables hacking of processors.

Actually, when they link back to publically disclosed security reports
or CVEs they kind of do. They provide enough detail for an experienced
programmer to reproduce the exploit. And that's what I'm asking you for
but not getting.

So far none of the vulnerabilities you've linked technical details on
have anything to do with PSN or TPM, so your assertion that PSN
"enables" these hacks is unfounded.

> Your question of how a website determine the geolocation of a
> client is a separate topic.

Excatly my point. Accurately determining the goelocation of a remote
node is NOT a trivial issue. So unless they have some solution and have
been holding out on the rest of the industry for the past decade, I'm
hesistant to believe that they have anything better than a geoip look

> Browsers, such as Firefox, have
> geolocation enabled. Most people do not know that there is an
> option to disable the geolocation in Firefox. Google Gears tracks
> geolocation offline. There are other Google apps that track
> geolocation which are used by websites tracking the geolocation of
> their visitors. So what UUIDs are Google apps using to track
> geolocation?

Again, these are cases of specific pieces of software making assumptions
about the geographic location based on clues provided by the user
(timezone, locale settings) and the public IP address (geoip lookup).
Not the most accurate/reliable, or even granular of information. It
typically isolates you to a rather larger metropolitan area, not even to
a particular zipcode.

> "Geolocation can be performed by associating a geographic location
> with the Internet Protocol (IP) address, MAC address, RFID,
> hardware embedded article/production number, embedded software
> number (such as UUID, Exif/IPTC/XMP or modern steganography),
> invoice, Wi-Fi connection location, or device GPS coordinates, or
> other, perhaps self-disclosed, information."
> http://www.privacyinfo.org/geoip

Right, but it's not a question of what's ultimately possible, it's a
question of what is pragmatically achievable. What is currently enabled
by the software on the device. This all a matter of the software
choosing to look at and report on these different values. If FB doesn't
include software that reports on these values, then the fact that they
potentially exist is moot.

> I should not have to have the burden to take the time to research
> how PSN, TPM and ARM's TrustZone are used. They exist to enable
> tracking of computers offline and online by websites. Websites sell
> user information. Malware tracks UUIDs.

I'm not asking you to research how PSN, TPM, and TrustZone are used. I'm
asking you to explain, in sufficient technical detail, the specific
claims you have made about how they are being used. You have made a
number of statements about these technologies which seem to be supported
by the available technical details of how they work. You are asserting
that having the hardware present is enough to create a vulnerability,
when according to all the credible documentation on how these chips work
and what they do the are essentially innert without the cooperation of
software running on the system which is explicitly making use of the
functionality they provide. It's beginning to sound to me like you might
not fully understand the technical implementation of these technologies.
You seem to be hung up on the worst case scenarios proposed by people
like RMS (who is well-known for embellishing the truth in his

> You do not need to know everything to ask Marvell whether their PSN
> is visible and whether there is ARM TrustZone in their motherboard.
> Please ask and disclose the answer on FreedomBox's website.

So then why don't _you_ ask them, and have them reply to the list or on
a public blog so that we can all hear the answer stright from the
horse's mouth?

Best regards,
Ben the Pyrate

Version: GnuPG v1.4.11 (GNU/Linux)


Reply to: