I like how negative things like this never makes it to the Front Page and not many throwing a punch at Apple like they do for other companies. I am sure if it was any other company, this news would have been the first frontage news. I feel this is done deliberately by MacRumors for generating $$. Just pathetic.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Changes in iOS 7 Security Make Kernel More Vulnerable to Attack
- Thread starter MacRumors
- Start date
- Sort by reaction score
I came across this story on the front page of MacRumors, as I'm sure many others did.
The new iPhone. We made everything thinner. Including security and the randomness of numbers.
It matches Ive's vision for Apple - form over function.
Just had a quick look - it appears that ARMv7 doesn't have it but ARMv8 does so I guess it is a matter of time before ARMv8 becomes the standard but until then I guess there always be a weakness somewhere in the system.
ARMv8 already is a standard. Apple makes their own CPUs, the A7 is an ARMv8 CPU.
They do use hardware for cryptography.
Last edited:
When I mean standard I mean that it is available on all their i-device products and it is the majority of the architecture out there.
If they do use hardware encryption then how do you explain the vulnerability in the original article? not being argumentative but just wondering.
When I mean standard I mean it's an established spec, which means someone can build something that implements the spec.
There may be areas where a cryptographically strong random number is needed and others where it's less critical. Maybe you are on to something, about not all iDevices running an A7. But since this is software it can be updated.
http://images.apple.com/iphone/business/docs/iOS_Security_Feb14.pdf
Last edited:
These folks say it could be more vulnerable but did they actually successfully exploit the flaw they found. If they have then I might be worried. But at this point it sounds like a possible issue only. A theoretical. Not really a solid fact
I hope you mean anit-virus ware?
No code is mistake free, those 3rd party mistakes just add to the pool of attacks.
You're just giving a third party a chance to profit from both their own and Apples mistakes. With no incentive to fix their own.
So they replaced one floored system where the code could be derived based on time to another that can only be cracked with bruit force guesses. So one is no more secure than the other. In other words its probably no more or less than it was before. Of course the tin foil hat brigade will have us all believe its a government conspiracy
It's not in the slides but I'd be curious to know how much brute force is required?
It reads like a restart would require calculation to start again?
Could an app be crafted inside the sandbox to not only gather enough info but to also then have enough time to process that info to get the information it needs to launch an attack without highlighting is presents.
Yes "security by obscurity" = bad. Yes, could be better.
Still if attack needs more than minutes of full throttle processing it goes to take some fairly careful crafting to hid it. Putting more in "Alert but not Alarmed" territory.
It's very hard to say how much of a problem there actually is. My understanding - which may be wrong - is that this random number generator is used at the very early stages while iOS is booting, and is then replaced with something a lot stronger. There's the claim that the random number sequence could be predicted, but then I wonder which non-Apple software would be running on the device at the early stages when this random number generator is in use. Quite possibly none at all.
It would be more like 7.1.1
Nothing is bullet proof, the "new" iOS is still in its infancy compared to everything up to ios6 so sure apple had security more tight back when.
Nothing is bullet proof, the "new" iOS is still in its infancy compared to everything up to ios6 so sure apple had security more tight back when.
Random Number Generation
I was part of a project many years ago where we used ethernet noise as part of our random number generation. With all the sensors on these new phones, you'd think it would be trivial to combine noise from environmental sensors to make nearly truly random values possible.
I was part of a project many years ago where we used ethernet noise as part of our random number generation. With all the sensors on these new phones, you'd think it would be trivial to combine noise from environmental sensors to make nearly truly random values possible.
The slides seem to suggest they could get enough info within the sandbox to reverse engineer the key a while after the boot. They don't say if they managed to do it with in sandbox. I mean if took any time then they need to be fairly careful on processor use as the system is designed to kill off apps that use a lot of resources with the user imputing regularly.
This is very interesting. I wonder if this is simply a side effect of (a) all of the under-the-hood changes in iOS 7 (as there were a ton of those alongside the new UI design) and/or (b) the tighter unification with OS X Mavericks (than iOS 6 had with Mountain Lion [despite the older pair being substantially more visually similar than the newer pair]). I'll bet that Apple will tighten this with iOS 8. Apple has lately fallen into the pattern of "brand new breakthrough thing/version of product" followed by "refinement update". This is clearly evidenced in the following pairs of products:
-Mac OS X 10.5 Leopard AND Mac OS X 10.6 Snow Leopard
-Mac OS X 10.7 Lion and OS X Mountain Lion
-iOS 5 and iOS 6 (excluding the whole Maps nonsense)
-iPhone x and iPhone xS (for the last three consecutive pairs of generations)
-First Generation iPad and iPad 2
-Third Generation iPad and Fourth Generation iPad
-iPad mini and iPad mini with Retina (though to much less of a degree than the other examples [in terms of refinement at least])
My guess is that iOS 8 will follow suit with iOS 7. I'm also hoping the same happens for Mavericks too.
They'll let the 4S run it; but it'll run similarly to the 3GS running 6.1.6 or the 4 running 7.1.x. My guess is that the 4S will be the only A5 device to run it. At least it ought to be. iOS 7 does drag a bit on the first generation iPad mini (and looks much worse on non-retina displays) and the 5th generation iPod touch.
Weird. 7.1 was just fine (if not a bit smoother) on my iPad Air and 7.1.1 didn't seem to do anything to break that trend.