Announcement

Collapse
No announcement yet.

Kernel Developers Work To Block NVIDIA "GPL Condom" Effort Around New NetGPU Code

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

  • Kernel Developers Work To Block NVIDIA "GPL Condom" Effort Around New NetGPU Code

    Phoronix: Kernel Developers Work To Block NVIDIA "GPL Condom" Effort Around New NetGPU Code

    Linux kernel developers are working on tightening up the access around GPL-only kernel symbols and kernel shims that interface with proprietary kernel modules. This latest work is being driven by code recently put out for improving the Linux networking code where NVIDIA's proprietary kernel driver would be the initial consumer...

    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
    So.. conform or die? Nope, not viral at all.

    Comment


    • #3
      When you have Greg saying “Okay, now you're just trolling us”, you know you fucked up.

      Comment


      • #4
        Granting the supposition that NVidia cannot release FL/OSS drivers without revealing trade secrets (which I would not, but for the sake of argument), what is the "right way" to do this?

        Comment


        • #5
          I'd find it great if any kernel point update deliberately broke any out of tree parasites. Some people either should learn the hard way or get lost.

          Comment


          • #6
            Surprised that someone senior at fb.com screws up that much.
            Seem to me he's too much accustomed with bsd than gpl.

            Comment


            • #7
              ( ° ͜ʖ͡°)╭∩╮

              Comment


              • #8
                Originally posted by starkruzr View Post
                Granting the supposition that NVidia cannot release FL/OSS drivers without revealing trade secrets (which I would not, but for the sake of argument), what is the "right way" to do this?
                How, do you think, did AMD solve that?
                Last edited by entropy; 03 August 2020, 02:47 PM.

                Comment


                • #9
                  As I read the discussion on the ML (which contains some borderline old Linus-level bashing BTW) the problem is that the code is written specifically to make it easy to hook it up to the API exposed by the nvidia blob and it is hacky as a result. A sane design would involve some modifications around the kernel, including the GPU drivers. Obviously, you cannot do this with the blob so the author had to devise a hacky solution which has no place in the kernel.

                  Comment


                  • #10
                    Originally posted by entropy View Post

                    How, do you think, did AMD solve that?
                    I don't know. AFAIK there is still a closed source AMD driver and an open source one.

                    Comment

                    Working...
                    X