Dubious Research Discovers Ryzen vulnerabilites

[smilingcrow]

Ryzen CPU's do have all of that on the CPU yes, the ASMedia Chipset is just an extension of features, Ryzen doesn't actually need it at all, it just means less USB's and PCIe lanes, infact their server chips have no Chipsets at all.
There is no ASMedia tech in the CPU its self, ASMedia is simply a stand alone Chipset...

I was aware that AMD used a 3rd party for assistance on the chipset but it wasn't clear to me whether ASMedia also helped them implement the same features that are also in the CPU.
It does seem slightly strange that they bring in a 3rd party to help with the chipset yet implemented the same features in the CPU themselves!
The one difference remaining between the platforms is that the ASMedia chipset is a core feature of most AM4 boards whereas it is an addition to the Intel chipset with Intel.

Anyway, this isn't the major issue for me amongst all this.
Am I correct in saying that some of these alleged exploits don't require physical access or a BIOS update but just remote administrative access?
Not that administrative access is at all trivial but considering how many systems are compromised these days the thought of a deeper level of compromise that is persistent is scary.
What a terrible year for internet security so far.

 

[Rroff]

I was aware that AMD used a 3rd party for assistance on the chipset but it wasn't clear to me whether ASMedia also helped them implement the same features that are also in the CPU.
It does seem slightly strange that they bring in a 3rd party to help with the chipset yet implemented the same features in the CPU themselves!
The one difference remaining between the platforms is that the ASMedia chipset is a core feature of most AM4 boards whereas it is an addition to the Intel chipset with Intel.

Anyway, this isn't the major issue for me amongst all this.
Am I correct in saying that some of these alleged exploits don't require physical access or a BIOS update but just remote administrative access?
Not that administrative access is at all trivial but considering how many systems are compromised these days the thought of a deeper level of compromise that is persistent is scary.
What a terrible year for internet security so far.

It seems unclear regarding what is on the CPU - they've changed and/or removed some of the information originally on the site in that respect and it now only talks about the chipset on the motherboard it would be a little strange if they used their own implementation on the CPU and a 3rd party for the chipset duplicating a lot of what they've already done but for instance the on-die controller is only USB3.0 while the external controller is USB3.1 so it might suggest that the on-die is entirely AMD's efforts.

None of these exploits absolutely depend on physical access and only the first one depends on a BIOS update supposedly according to the clarifications they published recently - they do all require elevated privileges running the code used however. The persistence is probably one of the more notable aspects of this and the only thing that is particularly concerning really as the rest depends heavily on having broken into the system already - however as I've illustrated that does happen from time to time and in combination with these issues if they are verified (bare in mind that at least two people with backgrounds in this and 10 years of traceable history in the industry so probably not connected to the people behind this have agreed that atleast in concept these aren't purely rubbish) it would take potential intrusions to another level in some scenarios.

 

[humbug]

I was aware that AMD used a 3rd party for assistance on the chipset but it wasn't clear to me whether ASMedia also helped them implement the same features that are also in the CPU.
It does seem slightly strange that they bring in a 3rd party to help with the chipset yet implemented the same features in the CPU themselves!
The one difference remaining between the platforms is that the ASMedia chipset is a core feature of most AM4 boards whereas it is an addition to the Intel chipset with Intel.

Anyway, this isn't the major issue for me amongst all this.
Am I correct in saying that some of these alleged exploits don't require physical access or a BIOS update but just remote administrative access?
Not that administrative access is at all trivial but considering how many systems are compromised these days the thought of a deeper level of compromise that is persistent is scary.
What a terrible year for internet security so far.

CTS Labs findings apply to Intel's ASmedia Chipsets as much as they do AMD's

The whole thing with CTS Labs findings is none of it is out of the ordinary for exploits on ANY platform, AMD, Intel or otherwaise, CTS Labs have simply taken what are very real vulnerabilities and conflated them as if they apply to AMD only, as if its a fault with the CPU's, when in fact this is no specific fault or in that way applicable to AMD, or Intel who have exactly the same vulnerabilities, Anand Tech put this to them, it was at this point CTS Labs ceased all communication with them.

We have to keep in mind the clear intention of CTS Labs, that is to cause as much damage to AMD as possible in order to devalue their stock so they can profit from them, that is what all this is about and apparent finds while not untrue are the same thing as downloading a Trojan to your computer and installing it, well duh......

 

[TrixP10]

It is sad some people think I'm just anti-AMD for the sake of it - notice that in the main 3 threads on Ryzen I have not posted at all - why is that? surely it would be prime estate if I just wanted to bad mouth AMD? (largely I just have nothing to contribute to those threads that hasn't already been said and unlike some are convinced I'm not out to rubbish AMD at every turn).

Everything I'm posting here is in good faith taking in a variety of sources and my own experience, which isn't without its flaws, but seems to be far more extensive than 9 out of 10 posters in this thread so far*. Infact other than Vince most of the posters here seem incapable of thinking for themselves and just regurgitating what they've read on the media articles and a couple of places like reddit without putting any thought into it and simply defending AMD out of hand until its proved otherwise, regardless of whether there potentially might be something to a particular angle or not, at which point if it is proved to be negative for AMD they usually slink off until the topic changes or just revert to making clown posts.

EDIT: Its also a bit sad you've obviously failed to read any of my posts in the light of the first two posts I made in this thread.

* Maybe one day they will actually prove me wrong with constructive and informative counter posts.

I am sure that you have posted in good faith and as I have said before your post are entertaining, possibly not very balanced or accurate but still entertaining for people who can recognize BS. The danger is that some people will believe what you or sites like CTS(cams) claim. Posting in good faith does not excuse you taking any opportunity to bad mouth a company based on just your opinion or unverified claims.
Everyone has an opinion. Good objective criticism is often well received. Personally, I think that within the timescale and resources that have been available to AMD they have done exceptionally well. Like most, there have been times when I have felt AMD have seriously underperformed and I have been more than just critical of AMD. At present I don't any reason to be critical of AMD. Possibly the people calling out AMD are just a wee bit jealous.

 

[CAT-THE-FIFTH]

Anand Tech put this to them, it was at this point CTS Labs ceased all communication with them.

Really??

 

[Rroff]

I am sure that you have posted in good faith and as I have said before your post are entertaining, possibly not very balanced or accurate but still entertaining for people who can recognize BS. The danger is that some people will believe what you or sites like CTS(cams) claim. Posting in good faith does not excuse you taking any opportunity to bad mouth a company based on just your opinion or unverified claims.
Everyone has an opinion. Good objective criticism is often well received. Personally, I think that within the timescale and resources that have been available to AMD they have done exceptionally well. Like most, there have been times when I have felt AMD have seriously underperformed and I have been more than just critical of AMD. At present I don't any reason to be critical of AMD. Possibly the people calling out AMD are just a wee bit jealous.

Very poor form on your part. Fortunately your post speaks more to your character than it does mine.

EDIT: I'll entertain you though - what have I actually claimed that is wrong? I've from the start emphasised that we still need to wait for AMD to verify a lot of this checks out, I've said most of these exploits require elevated privileges at which point all bets are off but I've also pointed out - and these have seemingly been independently verified as probably correct that there are some more concerning angles in certain environments. Also some of these issues are AMD specific although some aspects loosely mimic issues with Intel's AMT vulnerabilities (which as I showed I also posted about in a similar manner) their actual nature is different i.e. being able to work around memory protection at a low level which normally shouldn't be possible, etc. and the potential to use them to hide away malware in manners that isn't usually possible - generally Intel platforms aren't susceptible to the same exploits with the ASMedia controller although there are some boards that use them that might be - AMD has based their entire chipset around certain controllers as a standard while the same isn't true on the Intel side. Sure regardless of what system you are using being able to replace the BIOS with a customised one will give significant ability to compromise the system which isn't a "flaw" as such and not something that is an AMD problem or specific to AMD but I've not really touched on the Fallout part of these so called flaws as it pretty much goes without saying that the problems there aren't particularly different to allowing malware on your system, etc. but that is only one part of these flaws.

 

[smilingcrow]

CTS Labs findings apply to Intel's ASmedia Chipsets as much as they do AMD's.

If you continue to word it that way you will confuse newbies as there are no Intel ASMedia chipsets. There are premium Intel boards which use an ASMedia controller in addition to the Intel chipset but they are not integral so disabling it is relatively pain free.

The whole thing with CTS Labs findings is none of it is out of the ordinary for exploits on ANY platform, AMD, Intel or otherwaise, CTS Labs have simply taken what are very real vulnerabilities and conflated them as if they apply to AMD only, as if its a fault with the CPU's, when in fact this is no specific fault or in that way applicable to AMD, or Intel who have exactly the same vulnerabilities....

Knowing Intel's record over the last 12 months or so very little would surprise me in terms of their vulnerabilities.
But surely at the level of secure enclaves which presumably are not bound by x86 compatibility you would need a unique attack for each platform?
Or have security experts being stating categorically that Intel are definitely vulnerable and not theoretically but that proof of concept has been shown?
It's possible that Intel are less or more secure than AMD at this level but it seems too early to say at this stage.

 

[TrixP10]

Very poor form on your part. Fortunately your post speaks more to your character than it does mine.

Thank you. I am always grateful of character references

 

[CAT-THE-FIFTH]

Ok,I found this video on AT forums which was with Ian Cutress from AT talking with CTS-Labs:

https://www.youtube.com/watch?v=cj3_AILPvU0

A poster on AT forums who watched it said the following:

It's a long watch but there are some good details in there about the story and his call with CTS. Lots of questions he posed with no real answer, deflections, obvious lack of understanding of the modern server/compute environment, lack of understanding of modern security protocol, certain elements of their story changing, outright lying (according to other industry contacts) about not being able to share details with anandtech due to Israeli law, etc.

Ian is careful about not drawing complete conclusions which he shouldn't given his position and lack of expertise in all these areas, but luckily we don't have that standard in a casual tech forum and can call a spade a spade.

 

[Rroff]

If you continue to word it that way you will confuse newbies as there are no Intel ASMedia chipsets.

I don't think people have quite grasped that AMD has outsourced the external part of the Promontory chipset in a manner that hasn't typically been done in the past by either AMD or Intel which is one reason this is a little different to the normal vulnerabilities (assuming as always it checks out that these exploits are possible) - to my knowledge there are currently no known vulnerabilities in the Intel chipset that allow for the uploading of unsigned firmware for instance which is possible here. (EDIT: Quick google does find this however https://www.theregister.co.uk/2017/11/09/chipzilla_come_closer_closer_listen_dump_ime/ but it requires a physical connection to the board to accomplish).

 

[humbug]

@smilingcrow I'll get back to you when i have time on that.

But This is relevant to it, Roff every Chipset including Intel's provides access to update the microcode, if not directly and both Intel and AMD's are not volatile they use workarounds, its critical for fixes and updating.

 

[CAT-THE-FIFTH]

Transcript of AT interview with CTS-Labs:

https://www.anandtech.com/show/12536/our-interesting-call-with-cts-labs

Wow,they ended the interview early and didn't answer any more questions from AT!!

 

[gavinh87]

I am sure that you have posted in good faith and as I have said before your post are entertaining, possibly not very balanced or accurate but still entertaining for people who can recognize BS. The danger is that some people will believe what you or sites like CTS(cams) claim. Posting in good faith does not excuse you taking any opportunity to bad mouth a company based on just your opinion or unverified claims.
Everyone has an opinion. Good objective criticism is often well received. Personally, I think that within the timescale and resources that have been available to AMD they have done exceptionally well. Like most, there have been times when I have felt AMD have seriously underperformed and I have been more than just critical of AMD. At present I don't any reason to be critical of AMD. Possibly the people calling out AMD are just a wee bit jealous.

What are your opinions of vega?

 

[Hotwired]

The discussion of the phone call with CTS was interesting.

Didn't change anything about anyones position but there we go.

Comedy tweets from Viceroy desperately trying to sink AMD stock. AMD share prices meanwhile don't seem to care.

Also this from CTS:

https://safefirmware.com/CTO+Letter.pdf
https://safefirmware.com/Whitepaper+Clarification.pdf

Amazing stuff, they've decided the hyperbole on their site wasn't such a great plan and rewrote everything without the PR spin

So, still same position that there probably are vulnerabilities and the same suspect position of a share shorting company and security company getting together over it. Still needs more time.

What are your opinions of vega?

Really, a common private agenda rearing its ugly head to encourage a thread derail.

Hey lets stop this thread right here because I want to talk trash about graphics cards.

 

[Rroff]

Amazing stuff, they've decided the hyperbole on their site wasn't such a great plan and rewrote everything without the PR spin

Yeah they claim code snippets as well but I can't see them off hand - dunno why this information wasn't presented originally unless the whole thing was a sham from the start and they are now backfilling - but if that was the case assumedly AMD would have already released a statement to that effect.

Can see in the updated whitepaper the internal controller isn't shaded as a problem area while they've now highlighted which parts of the chipset are.

 

[gavinh87]

The discussion of the phone call with CTS was interesting.

Didn't change anything about anyones position but there we go.

Comedy tweets from Viceroy desperately trying to sink AMD stock. AMD share prices meanwhile don't seem to care.

Also this from CTS:

https://safefirmware.com/CTO+Letter.pdf
https://safefirmware.com/Whitepaper+Clarification.pdf

Amazing stuff, they've decided the hyperbole on their site wasn't such a great plan and rewrote everything without the PR spin

So, still same position that there probably are vulnerabilities and the same suspect position of a share shorting company and security company getting together over it. Still needs more time.



Really, a common private agenda rearing its ugly head to encourage a thread derail.

Hey lets stop this thread right here because I want to talk trash about graphics cards.

Merely trying to discover why you are not critical of AMD despite the crap storm that was vega.

 

[Hotwired]

Merely trying to discover why you are not critical of AMD despite the crap storm that was vega.

Well thanks I guess for the confirmation.

You see I'm not the person you originally tried to bait into a Vega argument, I am a different person accusing you of trying to derail a thread about a security firm and a share shorting firm apparently trying to leverage information they have about chip flaws.

...and off you go with the private agenda exactly as I accused you.

 

[TrixP10]

What are your opinions of vega?

Not for this thread - but

Over hyped and disappointed. My main concern was power consumption. Way behind Nvidia.
But then there is a lot more to vega than initially meets the eye. What we are seeing now is a better implementation. There are legs to it's design. APU's etc. (even Intel is buying them). I think it is probably a bit before it's time. Overly ambitious.
AMD know the basic design is good - but their professional customers want more for less power. They can make an awful lot of money in the professional market and I think that is the basis of Vega respin on 7nm. Time will tell. Mean time low power consumer / bitcoin is go. Gamers will have to wait - thats business.

 

[Emlyn_Dewar]

The Anandtech article is pretty interesting. It's nice that one of the tech sites is actually doing their job, and doing it damn well.

This story has highlighted just how trash some of the few sites I still read are.

 

[Rroff]

The Anandtech article is pretty interesting. It's nice that one of the tech sites is actually doing their job, and doing it damn well.

This story has highlighted just how trash some of the few sites I still read are.

Most sites these days just repost the same article with a few extra lines of their own - very few doing original work any more.