Announcement

Collapse
No announcement yet.

Dual monitors on fglrx 10.5 / Slackare 13.1

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

  • #11
    I think that in Adam?s case Xinerama is disabled already.

    Comment


    • #12
      oh sorry, I completely missed that pflynn wasn't the OP.

      in that case, the first two paragraphs of my last post go to adam while the third goes to pflynn.

      Comment


      • #13
        Well, I?m not sure, but maybe Adam could try to explicitly disable the XRandr extension in xorg.conf. Sometimes this extension conflicts with some resources in the Catalyst driver. But it is just a possibility...
        And in my case, I can?t disable de Xinerama extension, unfortunatelly, since It is needed in cases when one have a multi-card setup and one large desktop (ATI?s Big Desktop does not work in multi-card setup).

        Comment


        • #14
          Originally posted by pflynn View Post
          And in my case, I can?t disable de Xinerama extension, unfortunatelly, since It is needed in cases when one have a multi-card setup and one large desktop (ATI?s Big Desktop does not work in multi-card setup).
          I'm aware of that. Just saying that you should try to get working xorg.conf's for each of the cards, make sure each card is working fine on it's own. Once that's done, you can try joining them and add Xinerama. That'll help you narrow down the problems.

          No matter what, Xinerama is a pile of bugs. Prepare for pain and don't expect compositing to work.

          Btw, that segfault you posted seems to come from within X, not from fglrx.

          Comment


          • #15
            Originally posted by rohcQaH View Post
            I'm aware of that. Just saying that you should try to get working xorg.conf's for each of the cards, make sure each card is working fine on it's own. Once that's done, you can try joining them and add Xinerama. That'll help you narrow down the problems.
            Yes. Both cards work individually. If I configure my setup as

            Code:
            aticonfig --initial
            everything runs fine (but as a single, one monitor desktop). Also, the three monitor setup was working fine too, until the upgrade to 13.1 (using the same Catalyst 10.5).

            Originally posted by rohcQaH View Post
            No matter what, Xinerama is a pile of bugs. Prepare for pain and don't expect compositing to work.
            Yep. You are right. This is why I always disable compositing in my setup. By the way, compositing is not that important to me.

            Originally posted by rohcQaH View Post
            Btw, that segfault you posted seems to come from within X, not from fglrx.
            Agreed. This segfault seens to be related to X itself. And probably it has to do with, guess what? Composite I?m not sure, but I think composite is now enabled by default in the newer versions of Xorg. I?m trying to disable it in xorg.conf file, but I?m not sure if Xorg cares about it, since it look like it?s just ignoring my configuration.

            Comment


            • #16
              Originally posted by pflynn View Post
              Also, the three monitor setup was working fine too, until the upgrade to 13.1 (using the same Catalyst 10.5).
              so the driver version remains the same, but you're using a different version of Xorg? I have a strong suspicion about the real culprit here.

              It may be worth filing a bug-report. Prepare to be asked to re-create the bug with the OS drivers, blaming fglrx is always sooo easy

              Originally posted by pflynn View Post
              I?m trying to disable it in xorg.conf file, but I?m not sure if Xorg cares about it, since it look like it?s just ignoring my configuration.
              what makes you think the configuration is ignored? Any hints in Xorg.0.log?

              Comment


              • #17
                Originally posted by rohcQaH View Post
                so the driver version remains the same, but you're using a different version of Xorg? I have a strong suspicion about the real culprit here.

                It may be worth filing a bug-report. Prepare to be asked to re-create the bug with the OS drivers, blaming fglrx is always sooo easy
                Yes. I?m almost sure that the problem has to do with the new version of Xorg - not with Xorg itself, but between how fglrx and this new version are interacting. Downgrading the kernel, for example, probably will make no difference.

                Originally posted by rohcQaH View Post
                what makes you think the configuration is ignored? Any hints in Xorg.0.log?
                It?s just a guess. Because I have not found anything in the Xorg.0.log file showing that composite is being disabled when I explicitly write it in the xorg.cfg file (no messages like "disabling composite extension"). Also, window resizing now is allways very slow, no matter if I disable composition in the xorg.cfg file or not (when I?m using just one card). This behavior (window resizing being very slow) is kinda typical from fglrx working together with the composite extension. But I need to check it again to be sure.

                Comment


                • #18
                  Hi, good news.

                  I filled a bug report yesterday ( Bug 1828 - http://ati.cchtml.com/show_bug.cgi?id=1828 ), and today, an engineer from AMD, Felix Kuehling, replied to it:

                  Hi Pedro,

                  We know about the problem and a fix is in the pipe for release in July. In the
                  mean time you can try some workarounds:

                  - delete /dev/vga_arbiter before you start X, you'll have to do this again
                  every time you reboot
                  - if that doesn't work, try recompiling your kernel with CONFIG_VGA_ARB
                  disabled

                  Hope this helps.

                  Following the instructions did the job. Now I can start X successfully using multiple display adapters and Xinerama without segmentation faults. Many thanks to Felix and to everybody who helped me.

                  Comment


                  • #19
                    But of course, there is Adam's problem that is still pending. Now that I'm able to use my system normally, I'll see if I can help the owner of the thread.

                    Comment


                    • #20
                      I guess you missed the post where I said I managed to get it working :-)

                      Comment

                      Working...
                      X