Page 2 of 2 FirstFirst 12
Results 11 to 16 of 16

Thread: First Release Of The New Mode-Setting Driver

  1. #11
    Join Date
    Sep 2011
    Posts
    708

    Default

    Quote Originally Posted by uid313 View Post
    How is this any different from softpipe / llvmpipe?
    They are for 3d this is for 2d

  2. #12
    Join Date
    Dec 2007
    Posts
    2,404

    Default

    Quote Originally Posted by curaga View Post
    I'm curious, is there an advantage to running this over KMS FB compared to fbdev on KMS FB?
    Supports multiple displays which fbdev does not.

  3. #13
    Join Date
    May 2007
    Posts
    319

    Default

    Quote Originally Posted by uid313 View Post
    I am curious to know if this driver could be improved and how.

    Is it possible to add any type of acceleration to it?
    As far as I'm concerned its not the place to add acceleration. If you want acceleration, write a real driver for the hardware.

    Though I could probably be tempted to do something with the gallium XA state tracker, but it would depend on how ugly it would make things look.

    But yeah it's pretty much feature complete from what I'll be doing with it.

    the main reason is to have a driver for hw where I don't care about accel, but can work with multiple heads.
    Dave.

  4. #14
    Join Date
    Apr 2010
    Posts
    18

    Default

    Quote Originally Posted by airlied View Post
    As far as I'm concerned its not the place to add acceleration. If you want acceleration, write a real driver for the hardware.
    Out of interest, what are the main reasons not to aim at a generic X.Org modesetting driver for typical "modern" hardware? Maintaining three hardware-specific drivers seems a bit wasteful.

  5. #15
    Join Date
    May 2007
    Posts
    319

    Default

    Quote Originally Posted by michael-vb View Post
    Out of interest, what are the main reasons not to aim at a generic X.Org modesetting driver for typical "modern" hardware? Maintaining three hardware-specific drivers seems a bit wasteful.
    because optimal acceleration isn't generic, going forward maybe someone can use the gallium state tracker to make accel better but so far that hasn't happened.

    Dave.

  6. #16
    Join Date
    Apr 2010
    Posts
    18

    Default

    Quote Originally Posted by airlied View Post
    because optimal acceleration isn't generic, going forward maybe someone can use the gallium state tracker to make accel better but so far that hasn't happened.
    Certainly, but we already have the situation where the DDX driver and the KMS driver are working in tandem. Naively I would imagine that acceleration in the DDX driver could go through the DRI driver where that makes sense and be pushed down to the KMS driver where it doesn't. Not as in tomorrow of course, but in the longer term.

    On the other hand since I'm sure that other people must have thought of this too I assume that there are a couple of flies that I haven't spotted in that ointment.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •