Announcement

Collapse
No announcement yet.

NVIDIA's 302 Linux Driver Finally Has RandR 1.2/1.3

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

  • #16
    Originally posted by tuke81 View Post
    Not sure but those should be fixed according to nvidia driver team:
    http://www.nvnews.net/vbulletin/show...4&postcount=31
    Thanks a lot!

    Comment


    • #17
      Originally posted by Delgarde View Post
      Actually, no. You're right, it's an implementation detail, but it's a significant one. For the open drivers, KMS provides a single consistent method of mode setting, meaning that we have a solution that works for the console, for X, for Wayland, and anything else that might come up. It means that even lacking a decent hardware-accelerated driver, there's always the option to fall back to framebuffer.

      By not supporting KMS, Nvidia can't do that. You can't write a system that, e.g switches to a high-res mode for a bootsplash screen, then seamlessly hands over to X. Nor can you switch between X and a console without a mode switch. It can't even co-exist with the KMS support that the Nouveau guys have added. And that's because all that Nvidia provide is an X driver - a very good one, to be sure, but one which provides no integration with the standard way of doing things on Linux.
      Are those deficiencies really significant to the average user base though?

      I even get flicker during boot on Windows 7 (XP is far worse) and it's not really a popular complaint.

      Comment


      • #18
        Originally posted by Delgarde View Post
        By not supporting KMS, Nvidia can't do that. You can't write a system that, e.g switches to a high-res mode for a bootsplash screen, then seamlessly hands over to X. Nor can you switch between X and a console without a mode switch.
        Err, why not? Nvidia could write a fbcon driver that works on top of their modesetting which would do exactly that. There's no magic in KMS, it's just the X and the fbcon drivers cooperating. If Nvidia wrote a fbcon driver, they could also write such cooperation into it.

        Originally posted by johnc View Post
        Are those deficiencies really significant to the average user base though?
        I even get flicker during boot on Windows 7 (XP is far worse) and it's not really a popular complaint.
        This. And on my machine, I already get plenty of flickering in the boot process, before the machine even gets to Linux (hmm, does EFI fix that? I'm still in the BIOS world ). So one more flicker from console to X doesn't make any difference.
        Last edited by Gusar; 05-02-2012, 05:54 PM.

        Comment


        • #19
          Originally posted by Gusar View Post
          Err, why not? Nvidia could write a fbcon driver that works on top of their modesetting which would do exactly that. There's no magic in KMS, it's just the X and the fbcon drivers cooperating. If Nvidia wrote a fbcon driver, they could also write such cooperation into it.
          Why not? Because while Nvidia *could* do those things, they haven't. That's why not.

          I'm certainly not faulting them on the work they've done - for nearly a decade now, they've had the best supported graphics hardware on Linux, for those who don't mind binary drivers. It'd just be nice if they could adapt their work to fit the 'standard' way of working. Of course, I don't expect them to port their graphics driver to DRM, but the modesetting would be nice.

          Comment


          • #20
            Originally posted by Delgarde View Post
            Why not? Because while Nvidia *could* do those things, they haven't. That's why not.
            You were writing as if it was impossible for them to do so, as if there's a technical limitation or something. But they could do it, and they don't need KMS for it.

            Originally posted by Delgarde View Post
            I don't expect them to port their graphics driver to DRM, but the modesetting would be nice.
            But why? They don't need to, to provide the things people want when they say "we want KMS".

            xrandr is a bit different, it can do something twinview wasn't capable of - rotating just one display in a multi-display scenario. But besides that I'm not aware of any limitations twinview had, other than needing nvidia's tools to set it up, as opposed to using standard xrandr tools.

            Comment


            • #21
              Originally posted by asdx
              But still, no KMS, no plans to support Wayland...
              So when is wayland going to be stable? You know so there is something to target. I also thought that wayland could use Xorg drivers?

              How would KMS be helpful without an in-kernel driver?

              As much as I keep hoping nvidia open-sources their driver, it's not going to happen and that one that i can't spell or be bothered to look up, doesn't work so well on gentoo 64bit multilib (with 32bit wine) and doesn't really work well for gaming.

              Comment


              • #22
                I wonder if nvidia intends to add in some of the new gaming features that the Windows version picked up recently.

                Comment


                • #23
                  Originally posted by Gusar View Post
                  You were writing as if it was impossible for them to do so, as if there's a technical limitation or something. But they could do it, and they don't need KMS for it.

                  But why? They don't need to, to provide the things people want when they say "we want KMS".

                  xrandr is a bit different, it can do something twinview wasn't capable of - rotating just one display in a multi-display scenario. But besides that I'm not aware of any limitations twinview had, other than needing nvidia's tools to set it up, as opposed to using standard xrandr tools.
                  To add to that, well I'm pretty sure KMS is GPL, not LGPL, so there's likely licensing issues. As for another method, I doubt Nvidia wants to put in the effort for something non-critical. I'm sure if Linux becomes more popular, or Nvidia has some other incentive, then they'll probably do something about it.

                  I would concur that focusing on things such as critical issues, bugs and xrandr support seems like time better spent

                  Comment


                  • #24
                    Originally posted by Gusar View Post
                    You were writing as if it was impossible for them to do so, as if there's a technical limitation or something. But they could do it, and they don't need KMS for it.
                    Badly worded, perhaps. I wasn't saying it was impossible for Nvidia to fix - I was saying it was impossible to do certain things without them fixing it.

                    Comment


                    • #25
                      i wonder if this driver fixes the issue of powermizer being in full 3d clocks when 2 screens are operating.
                      when disabling one powermizer started to work normally.

                      Comment


                      • #26
                        Originally posted by xpander View Post
                        i wonder if this driver fixes the issue of powermizer being in full 3d clocks when 2 screens are operating.
                        when disabling one powermizer started to work normally.
                        I don't have that problem when both screens are the same resolution and refresh rate.

                        When my monitors were different resolutions, it was locked at the max. But that was the same in Windows 7 too.

                        Comment


                        • #27
                          Originally posted by johnc View Post
                          I wonder if nvidia intends to add in some of the new gaming features that the Windows version picked up recently.
                          like what?

                          I've opened a thread regarding ViewPortIn/Out here: http://www.nvnews.net/vbulletin/showthread.php?t=179892

                          Comment


                          • #28
                            Originally posted by xpander View Post
                            i wonder if this driver fixes the issue of powermizer being in full 3d clocks when 2 screens are operating.
                            when disabling one powermizer started to work normally.
                            That's a hardware issue. Generally, the memory clock cannot be easily changed without flickering if multiple screens are connected, and the driver defaults to the fastest clock due to that.

                            Comment


                            • #29
                              Originally posted by brent View Post
                              That's a hardware issue. Generally, the memory clock cannot be easily changed without flickering if multiple screens are connected, and the driver defaults to the fastest clock due to that.
                              Ah, so that's the reason? I always thought it's bandwidth issues. For example, on my laptop I have to force the card into it's lowest performance level, and when using TV-out (S-Video), the tv screen would go blank every few seconds. This wasn't the case when I could still operate the card in higher performance levels.

                              Comment


                              • #30
                                Originally posted by Gusar View Post
                                Ah, so that's the reason? I always thought it's bandwidth issues. For example, on my laptop I have to force the card into it's lowest performance level, and when using TV-out (S-Video), the tv screen would go blank every few seconds. This wasn't the case when I could still operate the card in higher performance levels.
                                The display controllers are isochronous clients of the memory controller, so you can not change the frequency of the memory while the dc does the scanout without distorting pixels on the screen. So to reclock memory you have to wait until the display is in VBlank and do the reclocking in this short timeslice. As soon as you connect two monitors, which always have a slightly different scanout rate (maybe just because of circuit tolerances) the two VBlank periods practically never sync up, so you have no chance to reclock memory without distorting the content of at least one screen. So the driver just goes to the highest mem clock and stays there, to work around this issue.

                                Comment

                                Working...
                                X