Friday, January 12th 2018
AMD Confirms They are Affected by Spectre, too
The public disclosure on January 3rd that multiple research teams had discovered security issues related to how modern microprocessors handle speculative execution has brought to the forefront the constant vigilance needed to protect and secure data. These threats seek to circumvent the microprocessor architecture controls that preserve secure data.
At AMD, security is our top priority and we are continually working to ensure the safety of our users as new risks arise. As a part of that vigilance, I wanted to update the community on our actions to address the situation.
Google Project Zero (GPZ) Variant 1 (Bounds Check Bypass or Spectre) is applicable to AMD processors.
We will provide further updates as appropriate on this site as AMD and the industry continue our collaborative work to develop mitigation solutions to protect users from these latest security threats.
Mark Papermaster,
Senior Vice President and Chief Technology Officer
At AMD, security is our top priority and we are continually working to ensure the safety of our users as new risks arise. As a part of that vigilance, I wanted to update the community on our actions to address the situation.
- We believe this threat can be contained with an operating system (OS) patch and we have been working with OS providers to address this issue.
- Microsoft is distributing patches for the majority of AMD systems now. We are working closely with them to correct an issue that paused the distribution of patches for some older AMD processors (AMD Opteron, Athlon and AMD Turion X2 Ultra families) earlier this week. We expect this issue to be corrected shortly and Microsoft should resume updates for these older processors by next week. For the latest details, please see Microsoft's website.
- Linux vendors are also rolling out patches across AMD products now.
- While we believe that AMD's processor architectures make it difficult to exploit Variant 2, we continue to work closely with the industry on this threat. We have defined additional steps through a combination of processor microcode updates and OS patches that we will make available to AMD customers and partners to further mitigate the threat.
- AMD will make optional microcode updates available to our customers and partners for Ryzen and EPYC processors starting this week. We expect to make updates available for our previous generation products over the coming weeks. These software updates will be provided by system providers and OS vendors; please check with your supplier for the latest information on the available option for your configuration and requirements.
- Linux vendors have begun to roll out OS patches for AMD systems, and we are working closely with Microsoft on the timing for distributing their patches. We are also engaging closely with the Linux community on development of "return trampoline" (Retpoline) software mitigations.
- We believe AMD processors are not susceptible due to our use of privilege level protections within paging architecture and no mitigation is required.
We will provide further updates as appropriate on this site as AMD and the industry continue our collaborative work to develop mitigation solutions to protect users from these latest security threats.
Mark Papermaster,
Senior Vice President and Chief Technology Officer
44 Comments on AMD Confirms They are Affected by Spectre, too
I'll repeat it once again: most media coverage (and comments on PC forums) are based on Project Zero's results. PZ created a method that was successful for Intel, i.e. it made it practically possible to leak data from the memory. This method was unsuccessful for AMD and ARM.
What happened next is very interesting. ARM confirmed that they succeeded with an attack, are affected and are working on a fix. AMD didn't. However, AMD's response evolved from being immune to:
"We believe AMD processors are not susceptible due to our use of privilege level protections within paging architecture and no mitigation is required."
A lot of people commenting on this are wrong saying that out-of-order execution doesn't work on AMD. It does. AMD says they have countermeasures that make it much harder to exploit.
Here's the original text from Project Zero. Check it out.
meltdownattack.com/meltdown.pdf
Here's the interesting chapter. I've underlined the fun part.
"6.4 Limitations on ARM and AMD
We also tried to reproduce the Meltdown bug on several ARM and AMD CPUs. However, we did not manage to successfully leak kernel memory with the attack described in Section 5, neither on ARM nor on AMD. The reasons for this can be manifold. First of all, our implementation might simply be too slow and a more optimized version might succeed. For instance, a more shallow out-of-order execution pipeline could tip the race condition towards against the data leakage. Similarly, if the processor lacks certain features, e.g., no re-order buffer, our current implementation might not be able to leak data. However, for both ARM and AMD, the toy example as described in Section 3 works reliably, indicating that out-of-order execution generally occurs and instructions past illegal memory accesses are also performed."Of course.
By far the best place to start is here:
www.intel.com/content/www/us/en/architecture-and-technology/facts-about-side-channel-analysis-and-intel-products.html
There's really a lot of stuff here which shows that most of the industry is really serious about this matter.
Intel's original whitepaper (fairly short one, but mentioning most important facts).
newsroom.intel.com/wp-content/uploads/sites/11/2018/01/Intel-Analysis-of-Speculative-Execution-Side-Channels.pdf
Following that, Intel posted list of CPUs affected and benchmarks.
security-center.intel.com/advisory.aspx?intelid=INTEL-SA-00088&languageid=en-fr
newsroom.intel.com/editorials/intel-security-issue-update-initial-performance-data-results-client-systems/
Next ARM site:
developer.arm.com/support/security-update
A lot of information, a beautiful, very detailed whitepaper, full list of CPUs affected and so on. Very nice stuff from the Brits - as usual. :)
By comparison, AMD's response is just a joke. There's nothing on the front page:
www.amd.com/en
The text we are discussing in this topic (www.amd.com/en/corporate/speculative-execution?sf178974629=1) is so shallow and pointless, that the reader won't even know what the threats are about. Even the CVE codes aren't mentioned!
I will be called an AMD hater again, but only thing I've been saying (since Zen launch) is that AMD is putting too much weight on gaming (and crypto mining lately...), so the whole product lineup and marketing around it started to look weird. They're sacrificing the rest of clients (including enterprise segment). A fast CPU is not enough!
The way they approached Spectre and Meltdown is another sign I'm right.
Look how outdated their EPYC website is.
www.amd.com/en/products/epyc
The latest news are from November. No mention of any Spectre problems.
ARM is also vulnerable (Nintendo Switch). Same is true (most likely / still somehow unofficially) for the IBM Power architecture (e.g. Xbox 360, Wii, PS3).Isn't it? Are you aware how people use consoles nowadays? Web browsers? Credit card information?
Console is a slightly limited PC. It has the same security problems.
Pff... even most modern TVs offer a web browser (and use an ARM chip).Yup. That's where power efficiency comes from... and the issues with high frequency... and potentially why they work best with Samsung RAM :p.
While I couldnt find something clear, most consoles seem to be Power5 or 6 based which isnt affected just like the current AMD consoles
And Nintendos Switch shouldnt be vulnerable too as its using a TX1. nvidia.custhelp.com/app/answers/detail/a_id/4616Do you have some reliable sources about that? Sounds a lot like fiction...:D
Design
- 32 KiB instruction + 32 KiB data L1 cache per core, L1 cache includes parity error detection
- 16-way, 1-2 MiB unified L2 cache shared by two or four cores, L2 cache is protected from errors by the use of error correcting code
- Out-of-order execution and speculative execution <-------
- Integrated memory controller
- Two-way integer execution
- Two-way 128-bit wide floating-point and packed integer execution
- Integer hardware divider
- Consumer processors support two DDR3L DIMMs in one channel at frequencies up to 1600 MHz[3]
- Server processors support two DDR3 DIMMS in one channel at frequencies up to 1600 MHz with ECC[4]
- As a SoC (not just an APU) it integrates Fusion controller hub
- Jaguar does not feature clustered multi-thread (CMT), meaning that execution resources are not shared between cores
The Switch uses A57 cores which are just as vulnerable as their other cores , at least according to ARM themselves.People tend to use consoles for a long time for multiple reasons. For example: unlike with PCs, you can't always use your games in the next one.
Also, many console gamers are not as hardcore as PC ones. Many of them casually play the same 5 games for a decade.
In 2016 PS3 and Xbox360 still had 10% of market share in console games revenue. I don't have any data for number of games sold or number of consoles in use. I think it's safe to assume that older games are cheaper and older console owners generally buy a lot less titles than those with latest models. Hence, the actual percentage of these 2 consoles in use should by much higher.Well... it's just a hypothesis without a proof. But there are signs. And it's not like I'm risking anything, while there is a nice prize possible: a huge "I told you so" on TPU. :p
Some time ago I got the feeling that Intel might want to buy AMD GPU part. They need it, they can afford it. some weird roadmap moves suggested a change of strategy. Sadly, Intel can't buy the CPU part (monopoly) and it's unlikely the CPU part itself would survive, since console sales are so important (and they need both CPU and GPU for that).
So this was just a guess - not really possible in AMD situation at that point.
Then we got some leaks about Intel CPU with AMD IGP. Then Koduri switched the companies right after Vega launch. Notice how quiet he is now - a guy that just few months ago was the most publicly active person in the industry. I never liked him - I don't care that much, but clearly he is busy right now. Then the Intel + Vega chip became a very lovely reality. (*)
In the meantime I've changed my opinion about AMD future from "selling Radeon" to "total collapse". And look... there's a potential buyer for the CPU part as well.
So if a 3-way deal was finalized, there would be no problem with anti-monopoly regulator. Intel Vega and Samsung Ryzen - in stores in 2019. :-D
Everything is fine, nothing to see here.
Thanks for the examples though, I appreciate it. I hadn't read those yet.Probably that's they've always been the best clocking ram for tight timings, and will accept damn near any timing the firmware configures.Complete fiction. The only similarity I am aware of is in the branch predictor, where they both use perceptrons. This is like saying that an orange and a potato are the same because they are both things that grow.
And up to 40% random 4k R/W performance for SSDs.
www.computerbase.de/2018-01/meltdown-spectre-amd-intel-benchmarks/#diagramm-star-wars-battlefront-2-1920-1080-intel-core-i5-2500k
So with CPU ram mother board and graphics cards prices being at a all time high.
Now this bug? ERROR comes out after someone knew for 10 years....
Now all we need is the malware / virus that attacks people.
Total screws windows 7 systems, older OS X
Have to have the latest Iphone / smart phones etc.
Pretty much makes so windows 10 and latest OS X are the only ways you will be safe....