Live data from Hacker News

The Intel ME subsystem can take over your machine, can't be audited

boingboing.net

201–210 of 282 posts

Re: The Intel ME subsystem can take over your machine, can't be audited

#201

Earlier quoted context omitted.

Indeed, the blob can be reverse engineered. Even more, an unbreakable signature can have it's private key stolen by hacking, by agencies inserting personnel into the companies, by agencies blackmailing key personnel and by agencies compelling the companies legally or ex-legally to hand them their keys.

Really, if someone has gone to the trouble of working out an exploit for Intel ME, the most ironic thing they could pull off would be to use that very exploit against Intel's own systems to steal their key, use it to patch the bugs, and release the patch to the world.

It'd be a spectacular successor to that router-patching virus that made the rounds a while back.

Re: The Intel ME subsystem can take over your machine, can't be audited

#202

Earlier quoted context omitted.

Wrong. 2048-bit RSA has an effective security level of approximately 112 bits, just like 3-DES. If you want 256 bits of security, you go all the way up to 15360-bit RSA. That's why Elliptic Curve systems are so attractive: they involve operations that are more costly per-bit, but the required key sizes to meet a security level are much less. Ofcourse, Elliptic Curve Cryptography as specified by NIST has its own downs…

Though when the NSA has done things like this in the past, we've found their choices prevented implementation weaknesses that weren't found (by anyone else) for several more years.

Can you follow up on that? I've never heard that story and I'm really curious.

On the narrower point, though, it's been shown that Dual_EC_DRBG is broken, and that the NSA values compromised the implementation instead of strengthening it.

Re: The Intel ME subsystem can take over your machine, can't be audited

#203

Earlier quoted context omitted.

Though when the NSA has done things like this in the past, we've found their choices prevented implementation weaknesses that weren't found (by anyone else) for several more years.

When has this happened? I'm curious about things that could cast the NSA in a positive light.

The DES standard's S-Boxes were changed by the NSA in the 1970s. It was long thought that this was to weaken them. However in the 1990's differential cryptanalysis was publicly discovered, and the NSA's changes to the S-boxes were found to have hardend DES agaist differential cryptanalysis.

Re: The Intel ME subsystem can take over your machine, can't be audited

#204

Amusingly, the ARC core in the Intel ME is a descendant of the SNES SuperFX chip.

I was quite tickled to read that bit. The guy who made Starglider for the Atari ST, and Star Fox! :-) Apparently newer chips have had other architectures, at least one I think was actually SPARC.

Re: The Intel ME subsystem can take over your machine, can't be audited

#205
post #42
post #14

Why can't Intel implement proper security and open up this blob to begin with? Not opening it and not allowing to disable it, suggests it's intended for something sinister.

As stated in the article, some researchers have managed to unpack it, and it can now be dissembled. You can't (and hpefully won't) be able to execute your own code there. There are 2 good reasons for this: 1) As per the article, to actually prevent ring -3 malware. The implemented signature is the best way to do this. If we could run our own "libre" code there, so could the attacker. 2) I bet this firmware controlls…

In many cases there will be manufacturing defects resulting in errors if you try to upgrade via firmware, especially early in a products lifetime when the defect rate is high. But of course there will still be many cases where you're able to upgrade without problems.

Re: The Intel ME subsystem can take over your machine, can't be audited

#206
post #61

Earlier quoted context omitted.

No, not that, the rest of it. I meant the fact that it's a binary blob which hasn't (recently) been subject to review by users.

The rest of it is covered by my second argument. They don't want you to see it due to highly proprietary stuff running there. And if you agree about code signing, do you really believe that letting everyone see it, but then not allow anyone else to change it, is a good idea?

It's better than not letting anyone see it, and also not allowing it to be changed.

At least in the case where the code can be inspected, it can be checked for vulnerabilities, backdoors, etc.

That is better than nothing, especially if the vendor can be asked for a fix.

Re: The Intel ME subsystem can take over your machine, can't be audited

#207

I'm very surprised that no-one on HN has talked about their experiences of using AMT for enterprise IT management. Aside from the security problems, I've personally never encountered or seen it's use, which makes the ME's inclusion (on all chips, for about 6 years) seem like an odd decision from Intel.

Its nice, I've seen demos of it. Its probably not in use a lot because there are so many better management suites out there. System Center on the expensive end, which is native MS, and a slew of other things on the lower end.

We just replaced our System Center setup and now use a mix of Spiceworks and PQDeploy, but we certainly could have looked at using AMT. Its such a buyers market and other apps have more features, it just didn't seem worth looking into. Intel's AMT/ME stuff seems more barebones than competitors. I can see why other shops are shying away from it.

FWIW, System Center was really, really nice and I wish we could have kept it, but MS discontinued the small office version of it and moving to the big boy version was just cost prohibitive. Unless you have over 250 machines to babysit, SC isn't worth it. SC eats a lot of competitors lunches. I suspect this is why big shops don't bother with AMT.

Also, AMT being unstoppable is a feature not a bug. You don't want end users being able to disable it or make changes if you manage an IT environment, even if they are local admins.

Re: The Intel ME subsystem can take over your machine, can't be audited

#208

This adds a whole new dimension to 'Intel Inside'. It says exactly what anyone needs to know. If it's for enterprise features as 'innocently' suggested that those who do not need or want this feature should be able to put it off simply without drama, debate or discussion. Its not surprising that both AMD and ARM have it. This is an orchestrated effort signifying the win of paranoia and security over privacy in the we…

This is simply a barrier of resources and technology. Let's consider how software became "free". An idealist, an university and a motivated Finnish student, among many others, were able to create two complete, free operating systems and toolchains, on top of which anybody and everybody in the world could build. Now free software is a resounding reality and even increasingly adopted by large corps who were 100% closed in the past.

The germane question is, can a similar revolution happen for hardware? Can motivated individuals, or small groups of people, reasonably hope to design AND manufacture ALL the hardware for a modern computer? The answer is it's quite beyond the bounds of possibility. The tech is too complex, too closely guarded and manufacturing has HUGE upfront costs.

THAT is why hardware is currently completely dominated by a few big players, which allows them (and any other "agencies") complete control to essentially do as they wish.

We were able to make software creation egalitarian. Unless we can do the same for hardware (from ground-up), we will be ultimately controlled and never be in full control.

Re: The Intel ME subsystem can take over your machine, can't be audited

#210

Igor Skochinsky (of IDA Hex-Rays fame, among others) has been studying Intel ME for quite some time. He gave a nice talk at Breakpoint summarizing what he'd discovered (slides here [pdf]: https://github.com/skochinsky/papers/blob/master/2014-10%20%... ). Among other things, he finds that ME is capable of running signed Java code which is pushed to the device. Due to the complexity and size of the Java code, it's quit…

ME is capable of running signed Java code How much firmware is in the thing? Is there a whole JVM in there? An OS? That's a lot of attack surface.

There is an embedded RTOS (ThreadX) and an embedded JVM. Curiously the embedded JVM uses JEFF bytecode, a somewhat obscure bytecode format for embedded JVMs. Java Card does _not_ use JEFF¹, meaning it's either an Intel-proprietary JVM or, more likely, they licensed it from an unknown party.

And yes, there's a hell of a lot of attack surface. Someone's gonna hack ME one day and have access to an awful lot of computers.

¹ I think. I'm far from an embedded Java expert, but from what I can tell Java Card uses a reduced form of regular Java bytecode and not a totally different format like JEFF. Please correct me if I'm wrong.

Post reply on HN