Announcement

Collapse
No announcement yet.

Differences Between X.Org, Wayland & Mir

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

  • #46
    Originally posted by erendorn View Post
    Still easier than rewriting closed graphic drivers.
    Canonical may ease that problem, and it will mean better gpu driver model on desktop.. They use Android gpu drivers just fine.

    And currently both Mir and Weston require rewrite of proprietary drivers.

    So its still true that both will require similar work.
    Last edited by przemoli; 03-26-2013, 11:33 AM.

    Comment


    • #47
      Originally posted by przemoli View Post
      Canonical may ease that problem, and it will mean better gpu driver model on desktop.. They use Android gpu drivers just fine.
      On Desktop the use libdrm, on mobile the use android drivers with libhybris. But this is nothing that you can't do with wayland.

      Comment


      • #48
        Originally posted by przemoli View Post
        Canonical may ease that problem, and it will mean better gpu driver model on desktop.. They use Android gpu drivers just fine.

        And currently both Mir and Weston require rewrite of proprietary drivers.

        So its still true that both will require similar work.
        That's the whole point. It will require twice the work, from outside the community, from corporations that are likely to be satisfied by doing the work only once. If you do a second init system and programs stop providing init scripts for upstart, canonical can just go on writing their init scripts themselves (they do it already). If you do a second display server and nvidia drops plans supporting the first, well good luck supporting their GPU on your own.
        That's why init systems, or packaging systems, or core tools, or filesystems, or network managers or much of the whole Linux stack cannot be compared to the display server, in terms of cost of forking/going non standard.

        Comment


        • #49
          Originally posted by erendorn View Post
          That's the whole point. It will require twice the work, from outside the community, from corporations that are likely to be satisfied by doing the work only once. If you do a second init system and programs stop providing init scripts for upstart, canonical can just go on writing their init scripts themselves (they do it already). If you do a second display server and nvidia drops plans supporting the first, well good luck supporting their GPU on your own.
          That's why init systems, or packaging systems, or core tools, or filesystems, or network managers or much of the whole Linux stack cannot be compared to the display server, in terms of cost of forking/going non standard.
          I'm not completely sure but I think this blog is about Upstart (from the original creator of Upstart) netsplit.com/2012/10/30/goodbye-ubuntu/ I get the impression different Init system require different code to support software raid somehow. And apparently none in Canonical care about the code in Ubuntu as its not there focus. Red hat has there own code and so has debian with sysvinit.

          Comment


          • #50
            Originally posted by erendorn View Post
            That's the whole point. It will require twice the work, from outside the community, from corporations that are likely to be satisfied by doing the work only once. If you do a second init system and programs stop providing init scripts for upstart, canonical can just go on writing their init scripts themselves (they do it already). If you do a second display server and nvidia drops plans supporting the first, well good luck supporting their GPU on your own.
            That's why init systems, or packaging systems, or core tools, or filesystems, or network managers or much of the whole Linux stack cannot be compared to the display server, in terms of cost of forking/going non standard.
            That assume that whole Linux ecosystem should just love what Wayalnd cooks with proprietary gpu vendors, but when its Canonical than nobody is allowed to touch it...

            No there will be just one new model for proprietary drivers. Wayland/Mir use it. However Wayland as a project was lacking leverage.. so it most probably be Mir team dictating what to change.

            Do note that what ever GPU vendors will come up with, it wont be some hidden stuff, everybody will be able to use it...

            Comment


            • #51
              Originally posted by erendorn View Post
              in terms of cost of forking/going non standard.
              Dude Wayland is non-standard and non-fork of X.org :P

              As its now, (almost) whole burden of additional work will fall on Canonical workers...

              Comment


              • #52
                Originally posted by johnny View Post
                no because nvidia will work on wayland .. there is absolutely 0 reason to believe otherwise.
                no because nvidia's open source driver will work on .. there is absolutely 0 reason to believe otherwise.
                There. Corrected for you.

                Comment

                Working...
                X