40 votes

Downfall security vulnerability in Intel processors

11 comments

  1. [3]
    skybrian
    Link
    This is a new attack allowing encryption keys and other sensitive info to be stolen by another process running on the same processor. Apparently there is a microcode update available from Intel. I...

    This is a new attack allowing encryption keys and other sensitive info to be stolen by another process running on the same processor. Apparently there is a microcode update available from Intel.

    I expect sysadmins are going to busy for a while?

    16 votes
    1. [2]
      Eji1700
      Link Parent
      I don’t envy them as this is just brutal. Affects a ton of Intel cpus, pain in the as to detect, has a patch that might butcher performance. Ugh. And it’s a major compromise

      I don’t envy them as this is just brutal. Affects a ton of Intel cpus, pain in the as to detect, has a patch that might butcher performance. Ugh.

      And it’s a major compromise

      8 votes
      1. Moonchild
        Link Parent
        It's highly unlikely to butcher the performance of any useful applications.

        It's highly unlikely to butcher the performance of any useful applications.

        1 vote
  2. [3]
    arghdos
    Link
    Between this and zenbleed, it seems like this sort of “leaked data between processes on the same core” vulnerability is picking up steam. Wonder if were going to see a whole class of them like Spectre

    Between this and zenbleed, it seems like this sort of “leaked data between processes on the same core” vulnerability is picking up steam. Wonder if were going to see a whole class of them like Spectre

    8 votes
    1. [2]
      spit-evil-olive-tips
      Link Parent
      we already have - this wiki article has a table. you can see the year as part of the assigned CVE number, there was a flurry of them back in 2017/2018, and then a relatively quiet period, and...

      Wonder if were going to see a whole class of them like Spectre

      we already have - this wiki article has a table. you can see the year as part of the assigned CVE number, there was a flurry of them back in 2017/2018, and then a relatively quiet period, and activity on them seems to be picking back up recently.

      there's been enough of them that lscpu on a Linux box will print out the vulnerabilities it knows about and whether there are any mitigations in place.

      from one of my laptops, for example, a Core i5-7200U which is from 2016 and so seems to require mitigations for a whole raft of them:

      Vulnerabilities:
        Itlb multihit:         KVM: Mitigation: VMX disabled
        L1tf:                  Mitigation; PTE Inversion; VMX conditional cache flushes, SMT vulnerable
        Mds:                   Mitigation; Clear CPU buffers; SMT vulnerable
        Meltdown:              Mitigation; PTI
        Mmio stale data:       Mitigation; Clear CPU buffers; SMT vulnerable
        Retbleed:              Mitigation; IBRS
        Spec store bypass:     Mitigation; Speculative Store Bypass disabled via prctl
        Spectre v1:            Mitigation; usercopy/swapgs barriers and __user pointer sanitization
        Spectre v2:            Mitigation; IBRS, IBPB conditional, STIBP conditional, RSB filling, PBRSB-eIBRS Not affected
        Srbds:                 Mitigation; Microcode
        Tsx async abort:       Not affected
      

      meanwhile, a much newer Ryzen 9 5900X requires mitigations for significantly fewer:

      Vulnerabilities:
        Itlb multihit:         Not affected
        L1tf:                  Not affected
        Mds:                   Not affected
        Meltdown:              Not affected
        Mmio stale data:       Not affected
        Retbleed:              Not affected
        Spec store bypass:     Mitigation; Speculative Store Bypass disabled via prctl and seccomp
        Spectre v1:            Mitigation; usercopy/swapgs barriers and __user pointer sanitization
        Spectre v2:            Mitigation; Retpolines, IBPB conditional, IBRS_FW, STIBP always-on, RSB filling, PBRSB-eIBRS Not affected
        Srbds:                 Not affected
        Tsx async abort:       Not affected
      
      12 votes
      1. arghdos
        Link Parent
        Huh, I really didn’t put together in my mind that zenbleed was really just another speculative execution bug like Spectre, but you’re totally right.

        Huh, I really didn’t put together in my mind that zenbleed was really just another speculative execution bug like Spectre, but you’re totally right.

        5 votes
  3. [4]
    Carrow
    Link
    Pardon my ignorance, but what would be the process for applying this sort of patch to my personal machine?

    Pardon my ignorance, but what would be the process for applying this sort of patch to my personal machine?

    6 votes
    1. [3]
      skybrian
      Link Parent
      I don't know. There is a git repo but practically speaking, maybe wait until there's an software update for whatever OS you're running?

      I don't know. There is a git repo but practically speaking, maybe wait until there's an software update for whatever OS you're running?

      6 votes
      1. [2]
        Carrow
        Link Parent
        Ah yes the practical solution was my interest. I wasn't sure if this sort of thing comes as an OS update or firmware update. Thanks!

        Ah yes the practical solution was my interest. I wasn't sure if this sort of thing comes as an OS update or firmware update. Thanks!

        2 votes
        1. em-dash
          Link Parent
          Specifically, microcode is a firmware update, but it's a type of firmware that's shipped with the OS and loaded during boot. You don't need to manually update it separately from the OS.

          Specifically, microcode is a firmware update, but it's a type of firmware that's shipped with the OS and loaded during boot. You don't need to manually update it separately from the OS.

          4 votes