Intel is literally selling garbage to their customers o_O

Ship me 5 of them for price of one and maybe then I'll be willing to consider them in product since I'll already have spare chips as they do what they are expected to do.

This always annoys me. After this long I should be used to it but I'm not. What is being called a stepping here is most likely NOT really one. It is the process revision. Intel nearly always uses a letter-number combination for the revision. The stepping is nearly always just a number. For example, on the i9 9900X I have the stepping is 4 and the revision M0. Another machine I have is a i7 3820 and it is on stepping 7, revision C2. My laptop is an i7 6700HQ and it is stepping 3, revision R0. These values are reported by CPU-Z. Revision and stepping are two different things. A new stepping usually means a new revision also but a new revision is not necessarily a new stepping. Stepping refers to the set of masks used by the lithography machines or the "steppers". A process revision encompasses all aspects of the fabrication process. There are around a hundred steps to process chips on a wafer and changing any of them constitutes a revision. Masks are rarely revised because it is a very expensive process, costing in the millions, which is why these chips have had so few of them. Again, refer to CPU-Z. It displays the stepping and revision and you can see that : 1) they are separate values, 2) the stepping is not in letter+number format - that is the revision, and 3) the stepping is just a number.

And nobody talking about this: Weakness in Intel chips lets researchers steal encrypted SSH keystrokes Intel Data-Direct I/O makes servers faster. It can also allow rogue servers to covertly steal data. https://arstechnica.com/information-technology/2019/09/weakness-in-intel-chips-lets-researchers-steal-encrypted-ssh-keystrokes/ Now, researchers are warning that, in certain scenarios, attackers can abuse DDIO to obtain keystrokes and possibly other types of sensitive data that flow through the memory of vulnerable servers. The most serious form of attack can take place in data centers and cloud environments that have both DDIO and remote direct memory access enabled to allow servers to exchange data. A server leased by a malicious hacker could abuse the vulnerability to attack other customers. To prove their point, the researchers devised an attack that allows a server to steal keystrokes typed into the protected SSH (or secure shell session) established between another server and an application server. Buuut,i9-9900K have integrated ze Data-Direct I/O support: https://www.intel.com/content/www/us/en/products/processors/core/i9-processors/i9-9900k.html

Last year I bought the AsRock J355B ITX with the intention of building my own NAS and a pFsense router. Well, that project somehow never realised fully, because of strange lockups and system freeze out of the blue. Looks like I won the bad lottery with those boards.

Turanis:

And nobody talking about this: Weakness in Intel chips lets researchers steal encrypted SSH keystrokes Intel Data-Direct I/O makes servers faster. It can also allow rogue servers to covertly steal data. https://arstechnica.com/information-technology/2019/09/weakness-in-intel-chips-lets-researchers-steal-encrypted-ssh-keystrokes/ Now, researchers are warning that, in certain scenarios, attackers can abuse DDIO to obtain keystrokes and possibly other types of sensitive data that flow through the memory of vulnerable servers. The most serious form of attack can take place in data centers and cloud environments that have both DDIO and remote direct memory access enabled to allow servers to exchange data. A server leased by a malicious hacker could abuse the vulnerability to attack other customers. To prove their point, the researchers devised an attack that allows a server to steal keystrokes typed into the protected SSH (or secure shell session) established between another server and an application server. Buuut,i9-9900K have integrated ze Data-Direct I/O support: https://www.intel.com/content/www/us/en/products/processors/core/i9-processors/i9-9900k.html

That's because it mostly affects servers. Unless I'm mistaken, it probably doesn't affect consumer products, which typically don't require high network IO.

I have one of these set up as a router (J3355)

anticupidon:

Last year I bought the AsRock J355B ITX with the intention of building my own NAS and a pFsense router. Well, that project somehow never realised fully, because of strange lockups and system freeze out of the blue. Looks like I won the bad lottery with those boards.

Yeah, I got one of those for a PfSense router as well.... I did get it working OK, mostly just to test PfSense. Was just about to get it back running again when I read this...

D3M1G0D:

That's because it mostly affects servers. Unless I'm mistaken, it probably doesn't affect consumer products, which typically don't require high network IO.

Have not read the article in other way than quick eye scan, but I expect that it applies to virtuals being able to get stuff from other virtuals. So, it would really not be any kind of problem for bare-metal systems. (Similar to AMD's vulnerability they patched in past.) And that's the question. Can intel patch it? Cloud is the future and this can be rather undesirable. And 2nd question is, where will this investigation lead. I think everyone remembers start of spectre/meltdown, and doors they did open over time. And those doors been reopen again after being closed...

Fox2232:

Have not read the article in other way than quick eye scan, but I expect that it applies to virtuals being able to get stuff from other virtuals. So, it would really not be any kind of problem for bare-metal systems. (Similar to AMD's vulnerability they patched in past.) And that's the question. Can intel patch it? Cloud is the future and this can be rather undesirable. And 2nd question is, where will this investigation lead. I think everyone remembers start of spectre/meltdown, and doors they did open over time. And those doors been reopen again after being closed...

I don't think VMs have much to do with it. From what I've read, the exploit concerns a shortcut that allows networked systems to communicate with each other at a much faster speed by using the CPU's cache instead of main memory. A compromised system that is hooked up to the same network can potentially steal data from other servers. A patch will probably mean disabling the feature (at a significant cost to performance) but I'm not sure if one will be forthcoming, due to the complex nature of the exploit - as long as systems admins carefully vet and qualify their servers, there should be little concern.

'In the revised notification, Intel appears to be making a clear distinction between the two most common use-cases for the parts: 'PC Usage' covers installation in low-end desktops, laptops, tablets and the like, which typically have a short design life before being replaced with a hardware refresh and which spend only part of their life actually powered on; the company's IOT Group, however, sells the same parts for Internet of Things embedded use in hardware which may have a decade-plus shelf life and run 24/7 - meaning it will meet whatever conditions trigger the degradation issue considerably more quickly than in the 'PC Usage' scenario.' Just sayin...while you might not agree with the binning or the sequence of events, they are two entirely different operating paradigms and use cases and would have different standards. No scandal, sorry.

they don't make em like they used to 🙄

HeavyHemi:

'In the revised notification, Intel appears to be making a clear distinction between the two most common use-cases for the parts: 'PC Usage' covers installation in low-end desktops, laptops, tablets and the like, which typically have a short design life before being replaced with a hardware refresh and which spend only part of their life actually powered on; the company's IOT Group, however, sells the same parts for Internet of Things embedded use in hardware which may have a decade-plus shelf life and run 24/7 - meaning it will meet whatever conditions trigger the degradation issue considerably more quickly than in the 'PC Usage' scenario.' Just sayin...while you might not agree with the binning or the sequence of events, they are two entirely different operating paradigms and use cases and would have different standards. No scandal, sorry.

Well maybe intel should put a value to that. "Not to be used 24/7" or something would have been nice. There was nothing like that on the Box.... Its an embedded CPU, not like it can be changed, its soldered.

Strange that a RUMOR about AMD Zen chips degrading spreads like wildfire. Yet an actual release from a manufacturer stating that their product has issues with degradation, goes largely unnoticed.