The embargo is still up until tomorrow. I expect there will be more info then.
Intel FUBAR ... again - Kernel memory leak in nearly every Intel CPU of the last decade (Spectre hits everyone, Meltdown still Intel exclusive)
Pholostan
#747
Meltdown patch in action gallery:
Five graphs with source links. All I’ve seen so far. All pretty bad.
Graphs collected by /u/ceremony64.
LinuxMaster9
#748
I included those to show how far back these problems have been. Most were not even aware of the AMT issues and we are trained to configure AMT on boards. I thought they should know.
Usually the spoiler does what the Details this does but it is now behind the Details.
catsay
#749
The sad reality of meltdown.
And why it makes most businesses no less secure than before 
torpcoms
#751
| CVE | Group | GPZ number | GPZ naming |
|---|---|---|---|
| CVE-2017-5715 | Spectre | variant 2 | branch target injection |
| CVE-2017-5753 | Spectre | variant 1 | bounds check bypass |
| CVE-2017-5754 | Meltdown | variant 3 | rogue data cache load |
Exploit homepages are spectreattack.com and meltdownattack.com, although both are the same.
Summaries so far
- @Noenken has a summary in his OP of this thread.
- I @torpcoms have a summary post in this thread that I still edit.
- @Freaksmacker was impressed by the summary from Carnegie Mellon’s SEI CERT (not US-CERT).
- Google has a quick summary, and a post on Project Zero’s blog.
- @catsay recommended Hector Martin’s summary/repo of helpful links on github.
- Bruce Schneier also has a link repository
- @noenken recommended the Raspberry Pi Blog’s post, which is pretty much Speculative Execution for Dummies. Not a summary, but really useful for explaining to others or yourself.
Here are some official pages from US Gov CERT or UK Gov CERT if that is more your thing. CERT-EU is just a glorified RSS/Atom reader, and doesn’t have a page.
@catsay also started a thread for trying to list CPUs that are not vulnerable at all, and I am trying to figure out what the situation is for POWER/PowerPC here.
Freaksmacker
#752
I have to agree…avoid the rag mag and get to the technical in the case of presenting something to your employer.
GeNerator
#753
Whilst i agree with the articles sentiment, it doesn’t stop the fact that an extremely hard to pull off remote exploit isn’t still serious.
catsay
#754
To use spectre you first need to get local code exec anyway. Usually there are far easier targets and attack vectors to achieve this.
Treat meltdown and spectre as more of a last resort system reconnaissance action. To extract tertiary information (passwords,ssh keys etc) that will then allow you to pivot further into a system / network.
GeNerator
#755
I know, i read the article. I agree with the sentiment but saying. Doesn’t stop it from being just as serious. I know, you know that. I can see it’s a call for calm but it still is serious. More and more things are going IoT they could run the most secure hardened OS going but if your physical hardware has built in exploits then were going to have a bad time.
catsay
#757
The S in IoT stands for security.
That said, I’m relatively positive about things. As long as we keep proding and poking into technologies as we did with what lead to this revelation. Then we are doing OK, we’re raising attention to this and bringing an awareness to the industry, an industry composed of many individual people.
People that have learned from this , that are now looking out for such problems more than ever before. We are already better off then we were a year ago. And while the path may seem long and littered with approaching problems, we shall tackle them each as we encounter them.
But no matter the odds we shall continue, we never give up or become complacent to it.
To a better future! May your hardware be trustworthy and your software free!
FurryJackman
#759
Not sure if it’s been posted already, but there are claims Intel released Coffee Lake anyways KNOWING this vulnerability existed:
catsay
#760
By that time the chips where essentially already made.
To fix it would be a disaster.
To silently release it would be less of a disaster.
Intel is a business.
Intel chose.
FurryJackman
#761
Still doesn’t protect them from anti trust lawsuits. It may just be mostly the US right now, but the EU might fine them hard. VERY HARD.
catsay
#763
They can go right ahead.
My prediction: It will be a slap on the wrist at best.
And the consumer will get the worst of it regardless. 
Also future prediction:
Intel will make a crazy redesigned security oriented CPU and then claim the high ground , that they have the most secure computing platform of anyone.
Then they will find issues with that CPU again
You’ve heard of Moore’s law.
But have you heard of the unwritten rule of computing?
There are always more problems.
sgtawesomesauce
#764
The problem is that intel knew about this 6 months ago. They put aside funds at that point for the legal fallout.
By the time these lawsuits close, 10 years from now, the fines will be a rounding error of a line item in their budget.
The problem isn’t so much the lawsuits/fines, but how long it takes them to execute. These companies need to be decimated every time they do shit like this.
I propose the CDP, Corporate Decimation Policy. Every time a large company makes a stupid move to cause lawsuits, the fine must be, at minimum, 10% of their earnings for the year.
GeNerator
#765
Can we also include putting the CEO and the board in stocks and throwing rotten vegetables at their face ?
j/k
sgtawesomesauce
#766
As much as that’s amusing, the 8th amendment of the US constitution prohibits cruel or unusual punishment, so no.