New Intel chip flaw threatens encryption, but Macs are safe

Posted:
in General Discussion edited March 2020
A chip-level flaw in Intel-made silicon could render many of the chipmaker's security features useless, though recent Macs are safe.

While the vulnerabiltiy affects all recent Intel processors, 10th-generation CPUs are safe from it.
While the vulnerability affects all recent Intel processors, 10th-generation CPUs and T1 and T2-equipped Macs are safe from it.


Inherent vulnerabilities in Intel chips have been a common theme over the past couple of years, with major flaws exploits like Meltdown, Spectre and ZombieLoad impacting virtually all Intel-equipped devices.

In 2019, security researchers at Positive Technologies discovered another issue with Intel chips. Specifically, it's a vulnerability affecting Intel's Converged Security Management Engine, a key security feature in Intel technology and firmware running on Intel hardware.

Along with loading and varying BIOS and power management firmware, CSME also provides the "cryptographic basis" for features such as Digital Rights Management (DRM) technologies, firmware-based trusted platform modules (TPMs), or Intel's own Enhanced Privacy ID.

Intel released a patch in 2019 to mitigate the issue, but researchers at Positive Technologies have found it to be much worse than originally anticipated. New research published on Thursday indicates that the vulnerability can be exploited to recover a root cryptographic key, potentially granting an attacker access to everything on a device's data.

That could be a major problem for DRM-protected media. Used offensively, the flaw could be leveraged to decrypt traffic inbound or outbound from the impacted device. On a larger scale, it could be used on Intel-based servers.

Though past Intel vulnerabilities have affected Apple devices, this flaw doesn't impact recent Macs equipped with an Apple T1 or T2 chip. Since those chips are based on first-party technology and boot before any Intel chips, a user's encryption keys are safe.

Of course, older Macs without a T-series chip -- or the present iMac lineup minus the iMac Pro -- may be vulnerable to the exploit, which could affect FileVault encryption. The flaw is unpatchable and Intel advises that users "maintain physical possession" of their devices as there is no way to use the attack vector remotely by clicking on a bad advertisement, for instance.

Intel points out that 10th-generation chips are safe from it, however. The vulnerability, and others like it, is also one of many potential reasons why Apple may soon move its Macs over to ARM-based processors.
«1

Comments

  • Reply 1 of 34
    razorpitrazorpit Posts: 1,796member
    FileVault protection on older Macs is useless? What do you say other than "Yikes"?
    darkvader
  • Reply 2 of 34
    Mike WuertheleMike Wuerthele Posts: 6,861administrator
    razorpit said:
    FileVault protection on older Macs is useless? What do you say other than "Yikes"?
    It's an issue, but not a giant day-to-day one. The timing attack is non-trivial to execute, is targeted, and if, like the article says, you maintain physical security of your Mac you'll be fine.
    llamalkruppcaladanianflyingdpdewmedysamoriaredgeminipaFileMakerFellerjony0watto_cobra
  • Reply 3 of 34
    rob53rob53 Posts: 3,251member
    razorpit said:
    FileVault protection on older Macs is useless? What do you say other than "Yikes"?
    It's an issue, but not a giant day-to-day one. The timing attack is non-trivial to execute, is targeted, and if, like the article says, you maintain physical security of your Mac you'll be fine.
    You don’t need FileVault if you’re protecting your equipment using physical security so your excuse for intel’s failure doesn’t hold water. FileVault is supposed to protect a Mac where physical security can’t be maintained so this revelation means intel, again, has failed to produce a secure CPU.  
    elijahgflyingdpsphericdarkvaderralphiedysamoriacaccamuccawatto_cobra
  • Reply 4 of 34
    SoliSoli Posts: 10,035member
    Yet another reason for Apple to ditch Intel for ARM as soon as it can, as the article states at the end.

    razorpit said:
    FileVault protection on older Macs is useless? What do you say other than "Yikes"?
    It's an issue, but not a giant day-to-day one. The timing attack is non-trivial to execute, is targeted, and if, like the article says, you maintain physical security of your Mac you'll be fine.
    rob53 said:
    razorpit said:
    FileVault protection on older Macs is useless? What do you say other than "Yikes"?
    It's an issue, but not a giant day-to-day one. The timing attack is non-trivial to execute, is targeted, and if, like the article says, you maintain physical security of your Mac you'll be fine.
    You don’t need FileVault if you’re protecting your equipment using physical security so your excuse for intel’s failure doesn’t hold water. FileVault is supposed to protect a Mac where physical security can’t be maintained so this revelation means intel, again, has failed to produce a secure CPU.  
    Sure, but FileVault exists because you can't always guarantee physical security of your Macs.
    edited March 2020 razorpitGG1watto_cobra
  • Reply 5 of 34
    Mike WuertheleMike Wuerthele Posts: 6,861administrator
    rob53 said:
    razorpit said:
    FileVault protection on older Macs is useless? What do you say other than "Yikes"?
    It's an issue, but not a giant day-to-day one. The timing attack is non-trivial to execute, is targeted, and if, like the article says, you maintain physical security of your Mac you'll be fine.
    You don’t need FileVault if you’re protecting your equipment using physical security so your excuse for intel’s failure doesn’t hold water. FileVault is supposed to protect a Mac where physical security can’t be maintained so this revelation means intel, again, has failed to produce a secure CPU.  
    It is ludicrous to assume that I'm making an "excuse for Intel's failure." The truth of the matter is that it is a hard exploit to take advantage of, for the discussed reasons.

    Dial back the rhetoric a bit.
    edited March 2020 gatorguykurai_kagellamablastdoorcaladanianflyingdpdewmeredgeminipajony0watto_cobra
  • Reply 6 of 34
    Whether this impacts FileVault depends on whether FileVault relies on "Intel CSME based encryption of data storage devices."  Does anyone know?  I wouldn't be surprised if the Apple solution is independent from Intel's encryption options (and therefore unaffected).
    edited March 2020 watto_cobra
  • Reply 7 of 34
    Mike WuertheleMike Wuerthele Posts: 6,861administrator
    Whether this impacts FileVault depends on whether FileVault relies on "Intel CSME based encryption of data storage devices."  Does anyone know?  I wouldn't be surprised if the Apple solution is independent from Intel's encryption options (and therefore unaffected).
    Early FileVault does. T1 or T2 implementations do not.
    randominternetpersoncaladaniandarkvaderdysamoriajony0watto_cobra
  • Reply 8 of 34
    razorpitrazorpit Posts: 1,796member
    razorpit said:
    FileVault protection on older Macs is useless? What do you say other than "Yikes"?
    It's an issue, but not a giant day-to-day one. The timing attack is non-trivial to execute, is targeted, and if, like the article says, you maintain physical security of your Mac you'll be fine.
    I’m thinking of things like Time Machine/USB drives when you travel. I’m generally ok when on the road, but we have people in our company that can’t walk from the office to their car without losing 2-3 things.  ;)

    And if you have a USB drive protected with a modern T2 Mac can it be broke by mounting it on an older machine?
    edited March 2020 watto_cobra
  • Reply 9 of 34
    linkmanlinkman Posts: 1,035member
    Is this a possible partial workaround? Use Filevault for full-disk encryption (just to make things a lot harder for the entity trying to access your data) but for your data save everything in an encrypted disk image -- and don't have Keychain save the password for it.
    watto_cobra
  • Reply 10 of 34
    wizard69wizard69 Posts: 13,377member
    Well this sucks!     I really hope that Apple gives AMD some design wins.   In most cases these days AMD offers vastly better performance at lower power levels.  Plus there are far fewer mitigation’s to deal with.  

    As for ARM based Macs I’d really love to see them.   Especially in laptop form.  I have this big fear though that Apple will lock them down tighter than an iPad.   In otherwords I love the thought of ARM technology but dress what Apple might do with it in a Mac.   
  • Reply 11 of 34
    SoliSoli Posts: 10,035member
    wizard69 said:
    Well this sucks!     I really hope that Apple gives AMD some design wins.   In most cases these days AMD offers vastly better performance at lower power levels.  Plus there are far fewer mitigation’s to deal with.  

    As for ARM based Macs I’d really love to see them.   Especially in laptop form.  I have this big fear though that Apple will lock them down tighter than an iPad.   In otherwords I love the thought of ARM technology but dress what Apple might do with it in a Mac.   
    1) I have no objection to AMD over Intel, but I don't see this ever happening if it hasn't happened yet.

    2) I would imagine it would be tighter than it is now, especially since we can see that the T-series chip locks it down more than before its existence, but I don't see it being tighter than the iPad. I'd say on par, at most.
    dysamoriawatto_cobra
  • Reply 12 of 34
    sflocalsflocal Posts: 6,095member
    This is just another nail in the coffin of Intel's relationship with Apple.  It's inexcusable.

    Like China, Apple also needs to sever its dependence on Intel.  AMD's tech is embarrassing Intel.  I prefer that Apple develop its own CPU's, but I do hope that some kind of x86(x64) compatibility is kept as many of us MacOS users do have to run Windows.

    The people responsible for these mistakes need to be fired.
    watto_cobra
  • Reply 13 of 34
    DAalsethDAalseth Posts: 2,783member
    sflocal said:
    The people responsible for these mistakes need to be fired.
    Then they and Intel won’t learn from this. It was an unforeseen exploit. Computers are full of them. If you can everyone who slips up, even makes a big mistake, soon you won’t have anyone working for you.
    dewmedysamoriathtStrangeDayswatto_cobra
  • Reply 14 of 34
    I'm not worried. I have nothing to hide.

    /s
  • Reply 15 of 34
    darkvaderdarkvader Posts: 1,146member
    The ARM Mac idiocy needs to die.

    The best thing Apple can do is start dual-sourcing processor chips - and that means Intel and AMD. 

    ARM chips are in no way suited to general purpose computer applications at this point, and probably never will be.  The ability to easily virtualize any common OS without having to emulate a processor is just too valuable to throw away on "Not Invented Here" stupidity.
    dysamoriamattinoz
  • Reply 16 of 34
    darkvaderdarkvader Posts: 1,146member
    mike54 said:
    It would most likely the vulnerability would be exploited by state actors, namely US/Israel, and at specific targets, whether it be individual, corporation or country. The average Joe is safe.

    Don't hold your breath.  This is a timing attack, not a crypto attack.  There are people who live for this sort of thing, I'd halfway expect a "popper" hardware box design to show up on the usual sources within a few months, and probably be buildable for under $100.

    Stolen MacBooks will be more valuable if they've got retrievable personal info on them.

  • Reply 17 of 34
    chasmchasm Posts: 3,303member
    You know what would be good? If Intel incorporated security into every level of the chip making, and not as an afterthought. You know, borrow a page from Apple’s philosophy on this.

    That’s not to say that it would eliminate all possibility of chip-level errors, but these stories about Intel’s chips would sure become less common if they took that approach.
    watto_cobra
  • Reply 18 of 34
    dewmedewme Posts: 5,368member
    DAalseth said:
    sflocal said:
    The people responsible for these mistakes need to be fired.
    Then they and Intel won’t learn from this. It was an unforeseen exploit. Computers are full of them. If you can everyone who slips up, even makes a big mistake, soon you won’t have anyone working for you.
    So true. But the more fundamental issue is that if you fire everyone who fails you will create a culture where nobody even tries. If this becomes the accepted norm, Intel would no longer exist, AMD would no longer exist, Apple would no longer exist, Microsoft would no longer exist, Google would no longer exist, Amazon would no longer exist, America would no longer exist, ... I hope you get the point.

    I hope you enjoy your abacus.
    sphericdysamoriaswat671chabigStrangeDayswatto_cobra
  • Reply 19 of 34
    CheeseFreezeCheeseFreeze Posts: 1,249member
    I hope T1 and T2 chips are being researched for exploits with the same attention as intel chips. It’s a smaller market and there’s no certainty Apple’s chips are without bugs.

    dysamoriaswat671
  • Reply 20 of 34
    knowitallknowitall Posts: 1,648member
    On a not so Positive note.
Sign In or Register to comment.