Crosstalk vulnerability in Intel processors allows information to be extracted from other cores
Click here to post a comment for Crosstalk vulnerability in Intel processors allows information to be extracted from other cores on our message forum
k3vst3r
This affect newer chips like 10900k?
Fox2232
Only important thing is performance loss. What @user1 wrote indicates that for certain instructions, loss is up to 97%.
asturur
I wanted to ask here exactly, what are those instructions doing? Because in this case would be better to patch compilers to use them as less as possible.
Fox2232
In that case Attacker will likely use his own tooling to maximize potential. When using default tooling will result in 100 times longer data extraction, it is easier to just use pre-patch compiler.
Goiur
New arq cant come soon enough for intel...
Undying
https://imagehost.imageupload.net/2020/06/11/159186270721914171.jpg
Glottiz
Who cares. None of these Intel "vulnerabilities" mean anything for end user. I have old intel CPU. Old mobo. No BIOS patches. The only security patches I have is what windows update provided. And guess what? I haven't had any issues, and I don't know anyone else who had.
sbacchetta
Intel is lucky that their server customer can't just jump to Epyc solution in the snap of a finger...
k3vst3r
Believe attacker can access crypto keys on Intel chips? meaning padlock symbol for SSH is useless whilst entering credit card information as one example. Hackers stealing your money no big deal?
Fox2232
Nobody will ever tell you story of how his SPI got out. Because they themselves will not know details even in case it was cyber attack.
kanenas
Tell that to 10 biggest data centers providers on the planet tell that to Oracle, Google,etc, atc even Nvidia now is using Epyc lol next 2 years gone by difficult for Intel at least for cpus.
asturur
This wasn't a solution for security, but for performance. You patch the CPU, but also for performance related application now you avoid those instructions if possible.
asturur
They mean for the end user when they can be exploited by a malicious software to overcome some anti-malicious software or when the patches gets embedded in windows and you loose performance.
The gamer is obviously looking at that, eventual performance loss that he wasn't planning to have.
I want to install the window patches as they come, i do not want to invest time in discarding the one that may affect performances.
Rich_Guy
https://thumbs.gfycat.com/FlakyCircularHamadryad-small.gif
JAMVA
The Cheese is getting mouldy Intel
o_O
Fender178
Well this is not good for Intel. I am glad that I am switching to AMD for my CPU. Well I guess it is safe to assume that CPUs below Broadwell are effected by this as well.
You are 100% wrong the end user would be the most likely effected group of this vulnerability if they don't know what they are doing.
schmidtbag
Kinda funny when you think about it: Zen2's rdrand was dysfunctional to the point that AMD's fix was basically to disable it. Thanks to security issues, Intel's rdrand is crippled so badly that it might as well be disabled.
I think we can effectively declare this instruction as obsolete.
I don't know how many times it needs to be said:
The entire reason these vulnerabilities are scary is because you can't detect them. You're not supposed to know if they're exploited; that's the whole point.
Kool64
And they all laughed at me when I said there were more
DeskStar
Oh they all are still..... Slowly, but surely.
DeskStar
Oh my.....oh my........OH MY..!.?!.?! When will this vulnerability train ever stop "chuffing" along?!?
