Skylake/Kaby Lake has new PortSmash side-channel vuln, possibly Zen too

[chipsnapper]

https://www.zdnet.com/article/intel-cpus-impacted-by-new-portsmash-side-channel-vulnerability/

"[PortSmash] definitely does not need root privileges," he said "Just user space."

"We leave as future work exploring the capabilities of PortSmash on other architectures featuring SMT, especially on AMD Ryzen systems," the research team said in a version of their paper shared with ZDNet, but Brumley told us via email that he strongly suspects that AMD CPUs are also impacted.

As if it wasn't before, Skylake is swiss cheese. If you do anything with any type of security you may need to consider moving platforms.

[EIREXE]

SMT is dumb

[ARitz Cracker]

Welp, there goes multithreading.

[frog]

Jfc Intel what are you even doing

[chipsnapper]

well the i7's already lined up with that fix

[nalyD]

The article is a little misleading at first, it takes a few paragraphs before it hints that this vulnerability may just be a fundamental issue with SMT and not a specific product issue. The researchers quotes seem to be really pushing that SMT is just bad

[Køleskab]

Christ! Is computing just going to get battered at every step from now on? Why is it that we went nearly 10 years with no one finding these issues and then we start getting all these found flaws in spectre, meltdown and now this? Who is figuring this shit out and why are they only finding it now? Is computational progress going to be dammed off like this every 18 months forever now?

Am i nuts for wondering who the fuck got access to computers these past 2 years that didn't have it before?

TL;DR: what the fuck is happening to computing right the fuck now?

[meefy]

Nobody bothered to look, likely because x86 is a shitheap. Anyway, I might actually have to set up my mips creator board and make that my daily driver...

[Demache]

I'm thinking that finding exploits in the way CPU's physically work is actually very difficult to do. And on top of that, creating a proof a concept showing its a real vulnerability.

This was found as part of a security research program. Its not like someone found this just goofing around. Its simply that nobody has been looking for these exploits. Or if they have, not made it publically known. Think NSA classified research.

[Glitchvid]

Nobody was really going after low-level x86 attacks for the most part (I mean, NSA, Unit 61398 , FSB, probably were).

This all underlines a significant issue with the closed, black-box approach CPU manufactures are going with, especially with co-processors like Intel ME, and AMD's PSP. Fundamentally, we need to move to an auditable and verifiable CPU security architecture, and having the co-processor be fully open and modifiable by the owner.

[BepisMax]

This really sucks it feels like any opportunistic latency absorption/hiding you do on a modern processor can open up a timing side channel.