NVIDIA Publishes PRIME Helper Patches

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts
  • phoronix
    Administrator
    • Jan 2007
    • 67141

    NVIDIA Publishes PRIME Helper Patches

    Phoronix: NVIDIA Publishes PRIME Helper Patches

    NVIDIA is still working on a way to implement buffer-sharing for their closed-source Linux graphics driver...

    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
  • dh04000
    Senior Member
    • Aug 2012
    • 892

    #2
    Does this solve the no nvidia blob access issue? Or is this something else entirely?

    Comment

    • blackout23
      Senior Member
      • Jul 2012
      • 1313

      #3
      And does it have any drawback to using DMA-Buff like initially intended?

      Comment

      • Bluefang
        Junior Member
        • Nov 2009
        • 7

        #4
        All this appears to do is consolidate code that was duplicated between the nouveau/radeon/exynos drivers.

        So the intro about the binary driver is a red herring.

        Comment

        • smorovic
          Junior Member
          • Nov 2008
          • 25

          #5
          Originally posted by asdx
          So they are still trying to be a leech and get away with their crap as usual?
          It looks as DRM interface to DMA-BUF, shared by drivers. Almost certainly it won't impose GPL requirement for usage, and so lets NVIDIA blob use DMA-BUF implicitly. It won't allow use cases of DMA-BUF with proprietary drivers, though. At least while devs remain stubborn

          Maybe NVIDIA should open their kernel part, like mobile GPU people did. Clever driver work is mostly done in userspace (although it would still be very useful to nouveau folks for easier RE).

          Comment

          • jonwil
            Senior Member
            • Oct 2008
            • 148

            #6
            I suspect a lot more is done in kernel space by the NVIDIA driver because of their need to share that driver codebase with OSX and Windows (on Windows in particular a lot more driver stuff is done in kernel AFAIK)

            Comment

            Working...
            X