Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Intel platforms from 2008 onwards have a remotely exploitable security hole

semiaccurate.com

Post reply on HN
Print view View original

101–110 of 190 posts

Re: Intel platforms from 2008 onwards have a remotely exploitable security hole

#101

Post by tomku »

Jan_jw wrote: post #94
tomku wrote: post #9

Is there a better source for this than SemiAccurate? The article doesn't really have much beyond self-aggrandizement and "we can't tell you any details, but you're screwed". For something that could be anything from "Charlie Demerjian heard a rumor about a ME patch and wanted some pageviews" to the actual security apocalypse, I'd like credible sources.

Yes. There is A better source: http://invisiblethingslab.com/resources/bh09dc/Attacking%20I...

That is not a source that confirms SemiAccurate's claims, it's a nearly decade-old paper (~2008-2009) describing an attack against a completely different security feature. The "System Management Mode" described and attacked in the paper is unrelated to Intel ME.

Re: Intel platforms from 2008 onwards have a remotely exploitable security hole

#102

Post by sixothree »

regularfry wrote: post #74
jackhack wrote: post #2

>>every Intel platform with AMT, ISM, and SBT from Nehalem in 2008 to Kaby Lake in 2017 has a remotely exploitable security hole in the ME (Management Engine) not CPU firmware. >>there is literally no Intel box made in the last 9+ years that isn’t at risk >>SemiAccurate has been begging Intel to fix this issue for literally years Am I the only one who is so cynical to think it must have been deliberate? Intel draggin…

Believe incompetence before malice, and I'd stick economic incentives somewhere in the middle. The discussion probably went something like: Person 1: "Should we issue a recall and disable a feature which bought us a several billion dollar customer?" Person 2: ...

In this day and age the choice between malice and incompetence seems to fall more on the malice side.

Re: Intel platforms from 2008 onwards have a remotely exploitable security hole

#104

Post by pmoriarty »

woodrowbarlow wrote: post #57
mschuster91 wrote: post #54

Earlier quoted context omitted.

Partially. Expansion cards use PCI-E which has DMA capability, so a bug/backdoor in their firmware can very well be used to attack a system. But I believe newer systems with MMUs acting as "firewalls" for DMA are safe from this vector.

there's also the concern of physical attacks, via the motherboard's RJ45 or USB.

A good argument to epoxy those ports shut, if you're really worried about that.

Re: Intel platforms from 2008 onwards have a remotely exploitable security hole

#106

Post by ripdog »

milcron wrote: post #97
Jan_jw wrote: post #94

Earlier quoted context omitted.

Yes. There is A better source: http://invisiblethingslab.com/resources/bh09dc/Attacking%20I...

What is the publication date of this?

>The details of our new SMM attacks will be made available once Intel patches its firmware, most likely we will present them at the Black Hat USA conference in summer 2009. We will also make the code of our TXT exploit available.

Re: Intel platforms from 2008 onwards have a remotely exploitable security hole

#107

Post by lurker456 »

LeifCarrotson wrote: post #23
lurker456 wrote: post #5

Great news that this finally came to light. After learning about remote management capabilities I've always suspected it had holes. Large attack surface, any exploit would have a high value, and closed source. Perhaps one day we'll be able to buy CPU's without this "feature". I'm betting AMD and ARM are in the same boat.

> After learning about remote management capabilities I've always suspected it had holes. Large attack surface, any exploit would have a high value, and closed source. Even after reading this, I'm still not convinced it does have holes. It's so high value (pervasive, incredibly powerful, and old) that if it were possible a bad actor would have used it . The spectrum of possibilities is small: 1. The hole does not exi…

why would 3 be unlikely ? The snowden leaks indicate that 3 letter agencies will use high value 0-days sparingly, to ensure they remain 0-days.

Also, there is a 5th (more likely) possibility: SA didn't find anything, but undiscovered holes do exist.

Re: Intel platforms from 2008 onwards have a remotely exploitable security hole

#108

Post by j_s »

mtgx wrote: post #18

Relevant discussion: https://news.ycombinator.com/item?id=11913379

"The Intel ME subsystem can take over your machine, can't be audited" - OP's discussion's title

> do the first three steps of thinking for them. Make it really easy for the other person to say yes or no

source: http://firstround.com/review/how-to-become-insanely-well-con... | https://news.ycombinator.com/item?id=14195664

I've agreed to include a bit of detail when spamming all my friends links via e-mail going forward.

Re: Intel platforms from 2008 onwards have a remotely exploitable security hole

#109

Post by eberkund »

davidgerard wrote: post #76
eberkund wrote: post #65

I've always wondering why nobody seems to notice the fact that this site is literally called "Semi Accurate". I mean sure, everyone makes mistake and even the most credible news sources are not entirely accurate all the time. But what am I to think when your organization is literally named after being only half truthful?

It's a semiconductor news site.

Semiconductor Accurate? Doesn't really sound right grammatically, also the arrows missing the target in their logo lead me to believe half accurate was how they intended the name to be interpreted.

Re: Intel platforms from 2008 onwards have a remotely exploitable security hole

#110

Post by wmf »

Sephr wrote: post #48

> For obvious reasons we couldn’t publish what we found It's not obvious to me why anyone not under an NSL or NDA would sit on this vulnerability for 5 years and wait until it's actively being exploited in the wild before public disclosure. It's extremely negligent to global security for SemiAccurate to not immediately publicly disclose the vulnerability 5 years ago after Intel refused to fix it. Of course this is ig…

One wonders if they knew of this particular vulnerability 5 years ago or they just knew that there must be vulnerabilities lurking in the ME somewhere.
Post reply on HN