Explicit GPU Synchronization Merged For XWayland

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts
  • dragon321
    Senior Member
    • May 2016
    • 873

    #11
    Originally posted by damentz View Post

    Not only is it worth the risk, but it'll bring immediate benefits that help current use cases. This quote from Eric:



    Pretty much if you're a power user of any type using more than one application at once, you benefit. Example, playing an intensive game on one display that dips below 60fps regularly, while watching a podcast on another. This is one of those last features that'll suddenly open up use cases that linux desktop really struggles with.
    That's right. There are good reasons why other modern windowing systems including Android (which is based on Linux) are based on explicit sync.

    Comment

    • Monsterovich
      Senior Member
      • Dec 2020
      • 298

      #12
      Originally posted by pharmasolin View Post
      Xorg fans, what else is wrong with a Wayland?
      Fragmentation, inferior architecture.

      Comment

      • skeevy420
        Senior Member
        • May 2017
        • 8644

        #13
        Originally posted by pharmasolin View Post
        Xorg fans, what else is wrong with a Wayland?
        That even though there's a reference implementation in Weston, everyone still does their own thing which includes adopting the various protocols at their own pace so we have a fragmented environment where it isn't guaranteed that one feature will be available from one graphical environment to the next.

        Because of that, for all the good Wayland does, it's fragmented by design and by default.

        That kind of design has its Pros. No one is beholden to a single organization or implementation; there's less lock-in. If kernels were done that way we could potentially swap out Linux for FreeBSD or Solaris. That's a hell of a pipe dream and a stretch of the imagination, but with more freedom comes fragmentation while defined protocols and targets comes unification and the potential for interoperability.

        It also allows the various teams to play Keeping Up With the Jonses. If KDE does all this nifty stuff and GNOME doesn't then GNOME starts to look bad and needs to keep up...or vice-a-versa, I'm not trying to play favorites here...because they're all playing Keep Up with wlroots and Weston

        Short-term, it sucks and makes Wayland look bad. Long-term, the Pros really outweigh the Cons.

        Wayland is sort of like Linux Distributions. You can pick the crappiest, worst distribution that exists and go "Welp, clearly Linux sucks." when there are 9000 other Linux implementations. KWin doesn't do Yada Blah so Wayland sucks. Let's ignore that Mutter or Sway does do Yada Blah.

        Comment

        • RejectModernity
          Senior Member
          • Dec 2021
          • 377

          #14
          Originally posted by Monsterovich View Post

          Fragmentation, inferior architecture.
          Nothing wrong with different implementations following stardards, whole world works like this. We do not need «Ein Volk, ein Reich, ein Führer»​. And what compared to is it inferior? To dead unmaintainable full of backdoors and crutches x11?

          Comment

          • ezst036
            Senior Member
            • Feb 2018
            • 680

            #15
            Originally posted by pharmasolin View Post
            Xorg fans, what else is wrong with a Wayland?
            Wayland or XWayland?

            Comment

            • MrCooper
              Senior Member
              • Aug 2008
              • 635

              #16
              Originally posted by M.Bahr View Post
              This merge request comes with a high risk to break things in the linux graphics stack and desktop environment.
              Not really. This MR itself carries very little risk as long as the client doesn't actually use the new PresentPixmapSynced request. And at this point there's no reason why Mesa must use it. So in the unlikely event that there's an issue, Mesa can just continue not using it until the issue is solved.​

              Originally posted by damentz View Post

              Not only is it worth the risk, but it'll bring immediate benefits that help current use cases. This quote from Eric:
              That quote from Erik is a bit misleading:

              What it describes can be done with implicit sync, per my blog post: https://blogs.gnome.org/shell-dev/20...nsive-clients/. https://gitlab.freedesktop.org/xorg/..._requests/1120 is an Xwayland implementation of this by yours truly. It was shot down by Nvidia engineers, citing performance concerns. Well, the explicit sync MR this article is about has the same performance characteristics. Apparently those can be acceptable or not depending on which synchronization colour they're painted in.

              Moreover, this can't actually have the described effect in Xwayland (in cases involving any Xwayland GPU work), because Xwayland's GPU context isn't higher priority than that of its clients.​

              Originally posted by schmidtbag View Post
              I'm a Wayland fan but from what I understand, multi-monitor seems to be the most problematic feature regardless of DE.
              Multi-monitor is one thing where Wayland has been pulling away from Xorg for some time.

              Comment

              • varikonniemi
                Senior Member
                • Jan 2012
                • 1102

                #17
                Originally posted by skeevy420 View Post

                That even though there's a reference implementation in Weston, everyone still does their own thing which includes adopting the various protocols at their own pace so we have a fragmented environment where it isn't guaranteed that one feature will be available from one graphical environment to the next.

                Because of that, for all the good Wayland does, it's fragmented by design and by default.

                That kind of design has its Pros. No one is beholden to a single organization or implementation; there's less lock-in. If kernels were done that way we could potentially swap out Linux for FreeBSD or Solaris. That's a hell of a pipe dream and a stretch of the imagination, but with more freedom comes fragmentation while defined protocols and targets comes unification and the potential for interoperability.

                It also allows the various teams to play Keeping Up With the Jonses. If KDE does all this nifty stuff and GNOME doesn't then GNOME starts to look bad and needs to keep up...or vice-a-versa, I'm not trying to play favorites here...because they're all playing Keep Up with wlroots and Weston

                Short-term, it sucks and makes Wayland look bad. Long-term, the Pros really outweigh the Cons.

                Wayland is sort of like Linux Distributions. You can pick the crappiest, worst distribution that exists and go "Welp, clearly Linux sucks." when there are 9000 other Linux implementations. KWin doesn't do Yada Blah so Wayland sucks. Let's ignore that Mutter or Sway does do Yada Blah.
                And this is it's largest benefit. No longer monoculture under one display server, but only a protocol to interface with the underlying stack, where diversity can flourish and the best will win! You call it fragmentation, i call it natural selection.

                And all fragmentation is only delusions in your head. Wayland is wayland and every wayland compositor is compatible. The fact some implement features not standardized is not fragmentation, it's just experimental work until a standard is agreed upon. Like this one in the article we talk about.
                Last edited by varikonniemi; 09 April 2024, 10:19 AM.

                Comment

                • skeevy420
                  Senior Member
                  • May 2017
                  • 8644

                  #18
                  Originally posted by varikonniemi View Post

                  And this is it's largest benefit. No longer monoculture under one display server, but only a protocol to interface with the underlying stack, where diversity can flourish and the best will win! You call it fragmentation, i call it natural selection.

                  And all fragmentation is only delusions in your head. Wayland is wayland and every wayland compositor is compatible. The fact some implement features not standardized is not fragmentation, it's just experimental work until a standard is agreed upon. Like this one in the article we talk about.
                  On Wayland you can't swap KWin for Mutter, XFWM, Fluxbox, IceWM, OpenBox, or another WM like we could back in the X11 days so they're not compatible like they used to be. Fragmentation via incompatibility. They just happen to support mostly the same Wayland protocols, but even now we can't switch between a GNOME, KDE, Mate, or Gamescope session and have the same features working or usable because that level of compatibility isn't part of the protocol.

                  Natural selection taking multiple years is short-term fragmentation and even after it takes place there's still that pesky fragmentation via the unswappable incompatibility that just isn't as prevalent with X11 due to all of them working off the single codebase.

                  Let's just agree to disagree on that.​

                  Am I the only one who rocked things like GNOME OpenBox and XFCE KWin back in the day?

                  Comment

                  • dpeterc
                    Phoronix Member
                    • Nov 2016
                    • 115

                    #19
                    Originally posted by M.Bahr View Post

                    I was not talking about the frame synchronization benefits of explicit sync in general but about the complexity of x11. The code is an ancient mess and just slight changes can break the DE experience. That's the reason why the wayland aka x11 devs weigh up the pros and cons and usually choose to not alter it. X11 is a mining field and nobody wants a broken desktop environment.
                    By the way explicit sync is in the making and planning by the mesa devs way longer before this merge request. I think the time and effort would have been better spent focusing on wayland and not on x11.​
                    Do you understand that the usual mantra of "X sucks and is an unmaintable mess" is quite out of place here?
                    The fix is for Wayland, not for X11.
                    It is not touching or fixing X11 problems, but the ones Wayland protocol failed to address properly.

                    Comment

                    • reba
                      Senior Member
                      • May 2020
                      • 676

                      #20
                      Originally posted by skeevy420 View Post

                      On Wayland you can't swap KWin for Mutter, XFWM, Fluxbox, IceWM, OpenBox, or another WM like we could back in the X11 days so they're not compatible like they used to be. Fragmentation via incompatibility. They just happen to support mostly the same Wayland protocols, but even now we can't switch between a GNOME, KDE, Mate, or Gamescope session and have the same features working or usable because that level of compatibility isn't part of the protocol.

                      Natural selection taking multiple years is short-term fragmentation and even after it takes place there's still that pesky fragmentation via the unswappable incompatibility that just isn't as prevalent with X11 due to all of them working off the single codebase.

                      Let's just agree to disagree on that.​

                      Am I the only one who rocked things like GNOME OpenBox and XFCE KWin back in the day?
                      Wayland is a standard and there are supposed to be multiple implementations of it.
                      Just like FTP clients understanding the file transfer protocol, even browsers were FTP clients; browsers themselves understanding HTTP, multiple systems using REST via HTTP and so on. There is no single implementation. And this is a very good thing. Chrome for all? GNOME for all? No alternatives allowed? GIMP? KCalc? Xorg?

                      Mixing and matching different desktops with window managers which are all based on Xorg; what for, they're all the same under Xorg. What's the use anyway? So, yes, I think your usecase was the corner case and I'm not sure everything worked with everything 100%
                      Last edited by reba; 09 April 2024, 11:01 AM.

                      Comment

                      Working...
                      X