Announcement

Collapse
No announcement yet.

AMD Posts Secure Memory Encryption For The Linux Kernel (SME)

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #11
    Same as Intel's:

    Originally posted by c117152 View Post
    It's DRM, not security. If security was it's purpose it wouldn't encrypt the physical memory. It would just deny read\write access.

    There is a legitimate case for memory encryption that has to do with privacy like Apple phones do it. But this isn't it.
    Sorry for quoting myself from that reply.

    Comment


    • #12
      Thanks.

      Comment


      • #13
        Originally posted by c117152 View Post
        It's DRM, not security. If security was it's purpose it wouldn't encrypt the physical memory. It would just deny read\write access.
        This is really server tech, not client (DRM is primarily a client issue).

        AFAIK the primary application of memory encryption is cloud computing, where a customer running on a VM doesn't want to have to trust the hypervisor or the hosting company. I imagine it could also be used for isolation between processes on a single OS instance, but it's aimed more at isolating VMs from each other and from the hypervisor.

        The other emerging need for this is going to be persistent memory now that is moving from "looking like a disk" to "looking like memory".
        Last edited by bridgman; 27 April 2016, 07:14 AM.
        Test signature

        Comment


        • #14
          Originally posted by Adarion View Post
          Michael, I highly DISAPPROVE of this approve-posts-thing.
          Michael disapproves of it too... it's an annoying bug, not a feature.
          Test signature

          Comment


          • #15
            Originally posted by bridgman View Post

            Michael disapproves of it too... it's an annoying bug, not a feature.

            what, comments have to be approved?

            Comment


            • #16
              AFAIK the spam-o-matic periodically decides that our posts might be spam and therefore need to be shunted to a moderator queue.

              Seems to have been getting worse recently... my guess is that the criteria for decision making is probably maintained by a third party (spam-o-matic provider) and that every time an update comes through our lives all become a little worse.
              Test signature

              Comment


              • #17
                Just me or does anyone else have to consciously think "Zen" is the AMD CPU, not the Hypervisor?

                So you can run Xen on Zen, with a virtual Zen on Xen on Zen.

                Comment


                • #18
                  Originally posted by bridgman View Post
                  AFAIK the spam-o-matic periodically decides that our posts might be spam and therefore need to be shunted to a moderator queue.

                  Seems to have been getting worse recently... my guess is that the criteria for decision making is probably maintained by a third party (spam-o-matic provider) and that every time an update comes through our lives all become a little worse.

                  ohh, yeah that sucks...

                  Comment


                  • #19
                    Originally posted by speculatrix View Post
                    Just me or does anyone else have to consciously think "Zen" is the AMD CPU, not the Hypervisor?
                    It's even worse for our internal emails, since the same team works on Xen support and Zen support. Summit Ridge (the platform name) is too long for convenient use, and if you abbreviate it to SR you get confusion with Steamroller.
                    Last edited by bridgman; 27 April 2016, 07:24 AM.
                    Test signature

                    Comment


                    • #20
                      I hope this new "feature" would not be an obstacle to reverse engineering.

                      Comment

                      Working...
                      X