Announcement

Collapse
No announcement yet.

Open-Source NVIDIA "Nouveau" DRM Changes Begin Queuing Ahead Of Linux 5.1

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

  • Open-Source NVIDIA "Nouveau" DRM Changes Begin Queuing Ahead Of Linux 5.1

    Phoronix: Open-Source NVIDIA "Nouveau" DRM Changes Begin Queuing Ahead Of Linux 5.1

    The Nouveau kernel driver tree where development happens on this open-source NVIDIA DRM driver saw a fresh batch of changes on Tuesday in aiming for new material with Linux 5.1...

    http://www.phoronix.com/scan.php?pag...1-Tree-Updated

  • #2
    Can we have a found raiser please and pay a developer to sift through the official Linux driver and its installer to find the necessary firmware required for reclocking?
    The data has to be there

    Comment


    • #3
      For sure it has to be there, but until the devs get NVidia's greenlight, they cannot legally distribute it with Nouveau. So there would still be hacks in the spirit of "please run this script on the NVidia installer to extract the firmware, bearing in mind that it was tested on v415.27 and may randomly break with any newer/older release".

      Comment


      • #4
        Originally posted by bemerk View Post
        Can we have a found raiser please and pay a developer to sift through the official Linux driver and its installer to find the necessary firmware required for reclocking?
        The data has to be there
        Wait, that's illegal.

        Comment


        • #5
          Originally posted by GreenByte View Post

          Wait, that's illegal.
          If part of the nouveau install process was downloading the nVidia drivers and extracting the firmware, that itself wouldn't be illegal.

          Comment


          • #6
            I think the situation is going to get better in the future.

            Intel and AMD already have great open-source support which is fully integrated with the standard Linux desktop stack (Mesa etc). Now that AMD GPUs are very competitive performance-wise with the nVIDIA ones, and Intel is entering this game as well, sooner or later nVIDIA will have to change its tune.

            Right now, we have to simply vote with our wallets. Unless you strictly need to use nVIDIA hardware - and the reasons for that are disappearing every day - buy from a vendor that really supports the open source stack. Also, tell your Windows-using friends which hardware vendor they should buy from. In the end money is the motivating factor for commercial entities.

            Comment


            • #7
              Originally posted by GreenByte View Post

              Wait, that's illegal.
              Just hire a russian dev and host it in a russian server.

              Comment


              • #8
                Originally posted by amehaye View Post
                I think the situation is going to get better in the future.

                Intel and AMD already have great open-source support which is fully integrated with the standard Linux desktop stack (Mesa etc). Now that AMD GPUs are very competitive performance-wise with the nVIDIA ones, and Intel is entering this game as well, sooner or later nVIDIA will have to change its tune.

                Right now, we have to simply vote with our wallets. Unless you strictly need to use nVIDIA hardware - and the reasons for that are disappearing every day - buy from a vendor that really supports the open source stack. Also, tell your Windows-using friends which hardware vendor they should buy from. In the end money is the motivating factor for commercial entities.
                I tend to disagree. Not entirely since I also think that "voting with your wallet" is the best message one can send to companies like Nvidia. Unfortunately, desktop linux GPUs are practically irrelevant to them, I bet. The real business with GPUs running on linux is compute farms and HPC. And apart from some efforts, which are just beginning to even come up with a plan, that is largely dominated by CUDA. Those HPC machines hardly ever run bleeding edge kernels since continuity and stability are the major requirements for them, which means that the binary driver is completely sufficient. Most of their users don't even care that with CUDA, they are not using an open standard and are tying themselves to one manufacturer. After all, they want to get their research done. IMHO, for Nvidia to start offering better open source support, the HPC crowd has to see some real benefits from it and demand it.

                Comment


                • #9
                  Originally posted by HadrienG View Post
                  For sure it has to be there, but until the devs get NVidia's greenlight, they cannot legally distribute it with Nouveau. So there would still be hacks in the spirit of "please run this script on the NVidia installer to extract the firmware, bearing in mind that it was tested on v415.27 and may randomly break with any newer/older release".
                  AFAIK, NVidia does provide older version of the binary drivers, back to 2011:
                  https://www.nvidia.com/object/linux-...y-archive.html

                  So picking one and using it should not be a problem. Actually a number of devices require running a script to obtain their firmware from binary/windows drivers. My PCTV dongle is one example, its script is part of the kernel source.

                  So DO IT.

                  Comment


                  • #10
                    I'm afraid it's harder than you believe. If I got it right, the "signed" firmware has to be "signed". I don't know the details, but I believe you can't just cut it out of a driver file and expect the GPU will load it. I understand that the driver must sign the damn blob once per GPU, maybe using a GPU public key and a NVidia private key or something like that. Not sure anyway, just guessing, but there must be a real reason for that (well deserved) finger...

                    Comment

                    Working...
                    X