Announcement

Collapse
No announcement yet.

Linux 5.5 To Advertise RDPRU Support For AMD Zen 2 CPUs Via /proc/cpuinfo

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

  • Linux 5.5 To Advertise RDPRU Support For AMD Zen 2 CPUs Via /proc/cpuinfo

    Phoronix: Linux 5.5 To Advertise RDPRU Support For AMD Zen 2 CPUs Via /proc/cpuinfo

    RDPRU is one of the new instruction set extensions of AMD "Zen 2" CPUs that is for reading a processor register that is typically limited to privilege level zero. RDPRU allows for reading select registers at any privilege level. With Linux 5.5, the RDPRU presence will be advertised by the CPU features...

    http://www.phoronix.com/scan.php?pag...ux-5.5-Feature

  • #2
    I foresee security issues in the future.

    Comment


    • #3
      Just out of curiosity.
      I wonder how many instructions a modern x86-machine has these days...? ISA extensions and subsets included.
      I kinda lost track of all variants and extra instructions quite a while ago.

      Anyone?

      Comment


      • #4
        Originally posted by milkylainen View Post
        Just out of curiosity.
        I wonder how many instructions a modern x86-machine has these days...? ISA extensions and subsets included.
        I kinda lost track of all variants and extra instructions quite a while ago.
        I think maybe once I saw a similar table like the one below but containing AVX, unfortunately I am unable to find it in a short time.

        (Source: https://net.cs.uni-bonn.de/wg/cs/staff/daniel-plohmann/)



        Register structure from Wikipedia with most registers up to AVX-512, although it seems K0-K7 opmask registers are missing from the picture:

        Comment


        • #5
          Originally posted by disobeyedtoast View Post
          I foresee security issues in the future.
          I foresee someone talking nonsense in the past, present and (unfortunately) future too. The RDPRU set is nothing different, security-wise and as a concept, than a setuid bit. Can it be used in a insecure fashion? Obviously yes. Is its existence to blame? Obviously not.

          Comment


          • #6
            Originally posted by milkylainen View Post
            Just out of curiosity.
            I wonder how many instructions a modern x86-machine has these days...? ISA extensions and subsets included.
            I kinda lost track of all variants and extra instructions quite a while ago.

            Anyone?
            I think you right..
            Too many bloat around.

            At least between [ 2034 - 3683 ]...but this number are higher for sure..
            Some details about some:
            https://www.felixcloutier.com/x86/

            Comment


            • #7
              This Youtube video might interest you. https://www.youtube.com/watch?v=KrksBdWcZgQ

              Comment


              • #8
                Originally posted by milkylainen View Post
                Just out of curiosity.
                I wonder how many instructions a modern x86-machine has these days...? ISA extensions and subsets included.
                I kinda lost track of all variants and extra instructions quite a while ago.
                Originally posted by tuxd3v View Post
                At least between [ 2034 - 3683 ]...but this number are higher for sure..
                Well, it is worth pointing out that a large chunks of those instructions are either:
                A. Long gone and obsoleted
                B. Only supported by one manufacturer (or in some cases, only one generation)
                C. Are directly related to each other (such as VPMOVQB, VPMOVQD, and VPMOVQW) or are an inverse of something else (such as EENTER or EEXIT).

                So - when you only look at widely-adopted currently-usable unique functions, it's not quite as bloated as it might seem. But, there are a whole lot of obscure instructions that I'm not sure will ever be used.

                Comment


                • #9
                  Originally posted by atomsymbol View Post

                  I think maybe once I saw a similar table like the one below but containing AVX, unfortunately I am unable to find it in a short time.

                  (Source: https://net.cs.uni-bonn.de/wg/cs/staff/daniel-plohmann/)
                  Thanks. Highly informational.

                  Comment

                  Working...
                  X