Announcement

Collapse
No announcement yet.

Intel 3.0 Linux Driver To Enable SNA, Support XMir

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

  • #16
    Originally posted by leech View Post
    I'm still angry at the whole 'Mir' thing. But besides that, a manager where I work who keeps pushing for me to use Ubuntu on servers instead of Debian didn't even know about Mir and the direction that Ubuntu is heading in with Unity, Ubuntu Touch, Ubuntu One, etc. When I told him about it, he had this weird, almost horrified look on his face. Then promptly shut up about promoting Ubuntu. Though his actual promotion of it was "You can just apt-get Oracle's Java7!" Uhm, no you can't...

    leech
    I'm not entirely sure the Mir thing is either relevant to your server install or even negative alltogether, let alone Ubuntu Touch...

    Comment


    • #17
      Originally posted by entropy View Post
      I guess this refers to this bug?

      https://bugs.freedesktop.org/show_bug.cgi?id=55500

      And there is no realistic hope that you'll eventually find a
      cure by means of an undocumented GPU setup?

      What is the usual way for the OTC developers in such a case:
      Do you directly talk to the intel CPU/GPU design team about those peculiarities?
      The hardware design and silicon validations teams are hard at work on the next generations, so we don't get any traction with old bugs unless we can demonstrate that they still affect the most recent generation (and preferrably the next one under active development). For old bugs like this, we can ask if anyone remembers something like it, but realistically we have to trawl through the list of known issues that were found by the validation and Windows driver teams and see if anything matches (and those are very terse descriptions with almost no explanation of the issue and often little detail as to the potential fix.) And no, I haven't found anything. At this moment in time, the only method I know that definitely works is to only let the GPU process one rectangle at a time.

      Comment


      • #18
        wait... what's this about rendering errors in gen4?

        back in the ubuntu 13.04 alphas when canonical was asking people to switch to SNA, I did and got only a few minor glitches in buttons and bars... I noticed that as the alphas progressed the glitches started getting more prevalent

        but SNA actually worked OK, once upon a time, in gen4 hardware as can be proven by this video:

        https://www.youtube.com/watch?v=xfN2sMJJU1s

        oh well, I guess that's another slew of old laptops that have been 'updated' out of linux... deja vu, exactly what happened to me with the idiots at xorg and the EXA debacle for old radeon igps.


        but don't be sad people you can always run windows 8.1 since the gen4 windows driver is not only much faster but also glitch free

        Comment


        • #19
          Originally posted by Pallidus View Post
          but don't be sad people you can always run windows 8.1 since the gen4 windows driver is not only much faster but also glitch free
          Actually, it is not.

          Comment


          • #20
            Originally posted by Vadi View Post
            To be fair, you can, that's how I get it on my desktop. webupd8's PPA for it is very stable and quite trustworthy after running for a couple of years now... at least for desktop use, anyway!

            How are desktop changes relevant to the server though, I've got no idea... You don't use a DE to manage your server, do you?
            Just running 'apt-get install oracle-j2sdk-1.7' and 'Add repository, apt-get install oracle-java7-installer then run installer script' isn't the same thing. In fact it is pretty much the same as what I had told him, though what I told him was "install java-package, download tar of java off of java.com, and run 'make-jpkg' on it, then it'll create a deb file, which you can install with dpkg -i.

            For the record as well, webupd8 also has the same thing for Debian, more or less.

            And ABSOLUTELY NOT, I don't use a DE on the server. Though he would love it if I did, and that is the main reason he wanted ubuntu on there, because he doesn't know how to use command line stuff to save his life. He uses vim instead of less to view logs...

            My reasoning for using Debian on a server over Ubuntu goes far beyond Mir, but I thought it was funny for some person who tries to advocate Ubuntu to not know about Mir at all.

            Comment


            • #21
              Originally posted by ickle View Post
              Actually, it is not.
              been using 8.1 since it first leaked and now on the official preview.

              whatever glitches/bugs there are I haven't been exposed to them and everything runs fine and smooth.


              using the driver from windows update that seems to be dated to last october
              (since intel decided that the official installer would no longer work even if you try to change the inf to recognize the system as win7)
              Last edited by Pallidus; 09-05-2013, 08:52 AM.

              Comment


              • #22
                Originally posted by r_a_trip View Post
                It's quite pragmatic actually. XMir support is higher priority, because Canonical is going to ship their next release with all DE's running on XMir.
                Nope, they won't.

                Comment


                • #23
                  Originally posted by Vim_User View Post
                  Nope, they won't.
                  Oh yes they will

                  Comment


                  • #24
                    All DE's, eh? So Enlightenment, XFCE, KDE, Gnome-Shell are all going to be using it? I can pretty much guarantee at least one of those won't be working.

                    Comment


                    • #25
                      Originally posted by leech View Post
                      All DE's, eh? So Enlightenment, XFCE, KDE, Gnome-Shell are all going to be using it? I can pretty much guarantee at least one of those won't be working.
                      Oh sorry missed the "all" word :P+

                      Comment


                      • #26
                        2.99.901 seems OK so far with SNA on my GM45.

                        Comment


                        • #27
                          http://cgit.freedesktop.org/xorg/dri...787d6b0fa64081

                          +Snapshot 2.99.902 (2013-08-07)
                          +==============================
                          +We do not condone or support Canonical in the course of action they have
                          +chosen, and will not carry XMir patches upstream.
                          +-The Management
                          +
                          http://cgit.freedesktop.org/xorg/dri...f9d304f23b188a

                          Revert "sna: Add XMir support"
                          This reverts commit 42d94356f65972eb7fb8991234a4e9388c4c2031. Ordered-by: The Management.

                          Comment


                          • #28
                            This should be fun: http://cgit.freedesktop.org/xorg/dri...f9d304f23b188a

                            Comment


                            • #29
                              Originally posted by ickle View Post
                              The hardware design and silicon validations teams are hard at work on the next generations, so we don't get any traction with old bugs unless we can demonstrate that they still affect the most recent generation (and preferrably the next one under active development). For old bugs like this, we can ask if anyone remembers something like it, but realistically we have to trawl through the list of known issues that were found by the validation and Windows driver teams and see if anything matches (and those are very terse descriptions with almost no explanation of the issue and often little detail as to the potential fix.) And no, I haven't found anything. At this moment in time, the only method I know that definitely works is to only let the GPU process one rectangle at a time.
                              Thanks for answering, Chris!

                              Originally posted by Gusar View Post
                              Uhh...

                              Serious (possibly stupid) question, is the license of a reverted patch still valid?
                              Or must it be treated as if it hasn't existed at all?

                              Also

                              Originally posted by NEWS, 2.99.902
                              Snapshot 2.99.902 (2013-08-07)
                              ==============================
                              We do not condone or support Canonical in the course of action they have
                              chosen, and will not carry XMir patches upstream.
                              -The Management
                              I guess the date should be 2013-09-07, but anyhow...
                              Last edited by entropy; 09-07-2013, 10:07 AM.

                              Comment


                              • #30
                                Originally posted by ickle View Post
                                It's just your GPU is broken. Sorry.
                                I tried using SNA for the first time a few days ago. Performance is overall *much* better, but sometimes there are some graphical glitches with text becoming stretched and slanted. Selecting the text or minimizing/maximizing the window will fix it.

                                Aside from those minor issues, SNA is great.

                                Comment

                                Working...
                                X