Intel Memory Access Design Flaw Already Addressed by Apple in macOS 10.13.2

Well that’s good. They’ve addressed it even before it was publicly available. My Hackintosh is happy!

 

I hope it's the same for Swift because the current main use for my Mac Pro's 12 cores is compiling my iOS client code in a reasonable amount of time.

 

offtopic: where can i find those wallpapers?

 

While ramping up the concurrent testing models for modular affectation, again using the Potts-Kant method -

Oddly enough, the iPhone X we were utilizing to tabulate and store the results caught on fire.

Luckily, none of the students in our lab here at Duke were assaulted - but the modular efficacy of the testing modules became subcutaneously lacrossed, as the fire consumed most but not all of the machines. This was not seen as a derisive fault of the iPhone, but - in honest, full foreclosure -

Potts-Kant continue to be under investigation here at the University, pending full results of our continued
floccinaucinihilipilification.

We'll be posting further results in another hour. We thank everyone for your continued up-votes and patience, and apologize for the delay.

 

Fortune article on the matter, AMD and ARM architectures are also vulnerable but details are still pending, AMD ensures their processors performance will not be affected, as for ARM... Well, we'll have to see what happens to apple's A series. http://fortune.com/2018/01/03/intel-kernel-security-flaw-amd/

 

Apple is impressive when it comes to rolling out security patches. Any word on Windows? Any word on Intel Chromebooks?

 

Chromebooks are basically Linux PC's, so it would depend on how fast google can roll out an OS with an updated kernel, there's a google official blog (just minutes ago) that includes ChromeOS info https://security.googleblog.com/2018/01/todays-cpu-vulnerability-what-you-need.html

 

Rhetorical question I am guessing but yes it could. You may consider returning it until things flush out more.

Unless you are made of money and can toss it when Apple ships systems with CPU’s that do not have this flaw I would be heading to the Apple store for a return.

This is going to crush the resell value of these system.

Investors must be thrilled as new system and cpu’s will be purchased to replace the millions of defective ones.

 

Hey Timmy ...... the craps going to hit the fan. We’re going to do you a solid and let you get ahead of this one. Remember that huge flaw we talked about a while back ... well ... we don’t think we can keep it under wraps anymore. Don’t worry about the NSA, we’ve contacted them and they want to thank both of us for allowing them to exploite it for this long, so don’t worry. Go ahead and patch it up, and you will come out smelling like roses. We owed you anyway for our chip delays, and we, of course, want to continue our business relationship with your fine company and would hope that this wouldn’t expedite any “conversations” you might be having with AMD or future plans at scrapping Intel processors for your own.

- Intel

 

Windows is patched on insider builds (and has been since November-ish), and should be released to GA this coming patch Tuesday.
Chromebooks are based off of linux, so should be fixed sooner than later.

 

I suspect in this case the exploit was discovered by Intel, which disclosed it to partners since it would require a software fix, but there was a NDA involved so the software engineers (for macOS, Windows and Linux) would have a chance to build a workaround to patch the vulnerability before the details of the vulnerability went public. Apple was just adhering to their legal requirements here. I do agree that the lack of transparency around power management in iOS was inexcusable though...

 

"partially addressed by Apple"

Wait for the fully addressed

But wtf, no update for Sierra & El Cap at least?

 

I don't think Intel wanted to make a general announcement about a major security vulnerability in their chips which can be exploited by hackers etc. before they gave their software partners had a chance to patch the vulnerability. It seems that info was leaked by someone ahead of the NDA being up though...

 

Ok, this is really bad news, but I have to say that I love when stuff like this happens because we get a chance to look at how the lowest levels of the system work.

To the extent that Macs seem to be handling this a bit better than others (preliminary, etc, etc), would it have anything to do with the hybrid XNU kernel MacOS uses?

 

So if this does affect ARM, then what about the elephant in the room? The huge number of Android devices that will never get updated to fix this issue?

And since Apple has their own microarchitecture, do A Series processors also have this issue? Or does it only affect those with processors based on actual ARM cores?

 

When will Apple fix it for Macs that cannot be upgraded to the latest OSX version?

 

While I don't know this, it's a fair guess that it was addressed in the security update for older systems, that came out at the same time. They do update the kernel-level stuff in these.

 

Well, I just upgraded from 10.12.6 to 10.13.2. I ran a few CPU-intensive benchmarks before and after.

Handbrake 1.0.7: 720p convert to 480p (fast): 2% loss in average FPS
Luxmark 3.1 Lobby (CPU C++): 3% decrease in score
Corona 1.3: ~0% change in average Rays/sec

The differences are so negligible that I would say there is no measurable change. I only ran each benchmark once due to time constraints.

This is on a flashed 4,1 Mac Pro running a Xeon x5677. I'll keep my notes and run these again if another patch is released.

 

I am running Sierra on my MacBook.Air and just checked for updates. I saw a "security update....10.12.6" listed. Plan to run it in the near future.

There's an update for El Capitan also.

 

And there may be other differences in the osx versions that could account for this too.

 

Microsoft announced that as of 5:00 PM EST the patch for Windows 10 users would be pushed automatically tonight. Other OS will be patched next week on the regular patch Tuesday cycle. Google to my knowledge has not said anything about it's plans, but should have has a leg up on everyone as it's researchers are the ones who discovered these bugs.

 

From my reading of the release notes for the security update, there is something in the kernel fixes that, while broadly worded, looks very much like this issue.

 

Nope they said they fixed, but not completely. We have no idea what this really means in Apple speak. Is it a 10% fix with 90% remaining or is it 90 fixed with 10 % remaining? Again no information that I have seen that gives us any idea.

 

this won't make me upgrade either.

Good, now all you gotta worry about is what happens after you do Apple updates One problem solved at least.

 

I expect that Intel won't be able to fix this issue until they move beyond the "Lake" series of designs. In other words, the upcoming Cannon Lake, Ice Lake, and Whiskey Lake will all likely have this problem. So - it will be years until there are fixed chips.