Announcement

Collapse
No announcement yet.

Intel Linux Kernel Graphics Driver Patched For New Security Sensitive Bug

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

  • Intel Linux Kernel Graphics Driver Patched For New Security Sensitive Bug

    Phoronix: Intel Linux Kernel Graphics Driver Patched For New Security Sensitive Bug

    CVE-2022-4139 was made public today as an i915 kernel graphics driver security issue affecting all Gen12 graphics -- from integrated Tigerlake graphics up through the latest Raptor Lake graphics as well as the in-development Meteor Lake code plus the discrete GPUs of DG2/Alchemist and Arctic Sound...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    Would this patch degrade graphics performance?

    Comment


    • #3
      Originally posted by bezirg View Post
      Would this patch degrade graphics performance?
      It doesn't look like it would. It's just changing a bit when the TLB is updated. At worst it adds a branch for the cpu to track, and potentially mispredict. So no performance change.

      Comment


      • #4
        Originally posted by xfcemint View Post
        ...
        You should mark that as sarcasm, in case someone takes you seriously.

        In case you are serious, you should read what the actual issue is. From my understanding, the bug is in how the driver is telling the GPU when it should no longer access a block of RAM. A mirokernel can't stop an iGPU from accessing arbitrary RAM.

        Comment


        • #5
          Michael

          Typo/grammar

          "as a incorrect" should be "as an incorrect"

          Lots of obviously missing commas

          e.g. "In some cases, "
          e.g. "At the very least, "
          etc.


          ‚Äč

          Comment


          • #6
            Originally posted by xfcemint View Post
            This kind of phenomena is absolutely unacceptable. It's all Linus' fault, since this would have much lesser impact if Linux was a microkernel.

            We can't expect the Linux ecosystem to advance while there remains such a high level of irrational stubbornes in the leaderhip of the project. Someone should go and wash the man's head.

            Too long article; DR.
            If Linus was so smart, why did he choose C? Why not a better language like Rust? Checkmate.

            Comment


            • #7
              Wonder if this affects FreeBSD and OpenBSD since they use the same Intel DRM, have to watch for a security update the next few days. Good thing it doesn't affect me personally since my *BSD rig is ZEN 2.

              Comment


              • #8
                Originally posted by kylew77 View Post
                Wonder if this affects FreeBSD and OpenBSD since they use the same Intel DRM, have to watch for a security update the next few days. Good thing it doesn't affect me personally since my *BSD rig is ZEN 2.
                I had the same thought. Only gen12 is affected, so there is a chance that they are unaffected by virtue of not having merged that code yet, but I have not checked.

                Edit: I just confirmed with a FreeBSD developer that FreeBSD is not affected. They have not imported gen12 support yet.
                Last edited by ryao; 01 December 2022, 11:06 AM.

                Comment


                • #9
                  Originally posted by xfcemint View Post
                  It doesn't matter what the actual issue is. I don't need to read the article, it is sufficient that I have red the title. The issue described in the article simply isn't an issue in a microkernel.
                  I mean, that might be true if your microkernel had no support for the intel GPU in the first place ;-)

                  The issue is about the management of the GPU's MMU/TLB, microkernel vs whatever else has no bearing on the issue

                  Comment


                  • #10
                    Originally posted by xfcemint View Post

                    You are wrong.

                    The issue is that you don't know the correct answer; more precisely: you can't figure out how a microkernel can help in this scenario of security bugs in a GPU driver.
                    No, actually I do know what I'm talking about here.

                    Comment

                    Working...
                    X