Announcement

Collapse
No announcement yet.

Linux 6.8 To Add Atomic Mode-Setting Mouse Hotspots, Atomic Async Page Flip

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

  • Linux 6.8 To Add Atomic Mode-Setting Mouse Hotspots, Atomic Async Page Flip

    Phoronix: Linux 6.8 To Add Atomic Mode-Setting Mouse Hotspots, Atomic Async Page Flip

    The drm-misc-next changes sent out this week to DRM-Next ahead of the Linux 6.8 kernel cycle include a few interesting additions worth mentioning for Linux desktop users...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    I wonder if the async function DRM_CLIENT_CAP_CURSOR_PLANE_HOTSPOT could aid in resolving cursor stuttering in Sway with VRR (and pretty much any wlroots based compositor)?

    Ref: https://github.com/swaywm/sway/issues/6832

    Comment


    • #3
      I was wondering the same (always do when atomic DRM API comes up), but this doesn't look like it. I'm pretty sure at this point that cursor stutter/lag is working as intended, perfect frames with no possible cursor tearing (never seen it myself even after trying hard...) at the cost of smooth, lag-free movement is a good tradeoff for the modern Linux desktop. Only solution is to just keep using the older non-atomic API. Didn't sway support that anyway?

      EDIT: Oh, I don't really know if that helps with VRR, though. That's its own can of worms until monitors start supporting cursor planes natively.
      Last edited by binarybanana; 03 December 2023, 07:39 AM.

      Comment


      • #4
        I don't know why all Wayland compositors have problems with mouse cursor. Every Wayland compositor I tested has cursor input lag (which varies depends on compositor and settings) and cursor stuttering with heavy GPU load. Also noticed in every Wayland compositor the FreeSync issues when moving the cursor.

        Xorg with modesetting (and amdgpu and Nvidia proprietary) drivers doesn't have any of these issues, so for me it doesn't looks like a missing kernel feature, but problems with all Wayland compositors implementation.

        The problem is that we have fragmentation in Wayland implementations - a lot of compositors and/or libraries written from scratch instead of one rock-solid implementation with all features which will be used by all desktop environments.

        Comment


        • #5
          So it will be even better to run steamos in a VM?

          Comment


          • #6
            Originally posted by zaps166 View Post
            I don't know why all Wayland compositors have problems with mouse cursor. Every Wayland compositor I tested has cursor input lag (which varies depends on compositor and settings) and cursor stuttering with heavy GPU load. Also noticed in every Wayland compositor the FreeSync issues when moving the cursor.

            Xorg with modesetting (and amdgpu and Nvidia proprietary) drivers doesn't have any of these issues, so for me it doesn't looks like a missing kernel feature, but problems with all Wayland compositors implementation.

            The problem is that we have fragmentation in Wayland implementations - a lot of compositors and/or libraries written from scratch instead of one rock-solid implementation with all features which will be used by all desktop environments.
            I think it might have to do with software cursor instead of hardware cursor.

            Comment


            • #7
              Originally posted by zaps166 View Post
              I don't know why all Wayland compositors have problems with mouse cursor. Every Wayland compositor I tested has cursor input lag (which varies depends on compositor and settings) and cursor stuttering with heavy GPU load. Also noticed in every Wayland compositor the FreeSync issues when moving the cursor.

              Xorg with modesetting (and amdgpu and Nvidia proprietary) drivers doesn't have any of these issues, so for me it doesn't looks like a missing kernel feature, but problems with all Wayland compositors implementation.

              The problem is that we have fragmentation in Wayland implementations - a lot of compositors and/or libraries written from scratch instead of one rock-solid implementation with all features which will be used by all desktop environments.
              'a lot of compositors and/or libraries written from scratch', but why do that programmers and software get into same difficulty and what compositor(/display manager?) implementation is most functional (especially with Wayland being only option with some distros and programs)?

              Is there an established environment variable (setting) for cpu supported rendering for compositors/display managers in Wayland (instead of partly lagging gpu hw rendering, with a cursor only being one visible sign for other rendering constraints)?

              A hardware cursor is an overlay picture for a cursor (style?), drawn/rendered from gpu into screen frame buffer(?)

              (Thx)

              Comment

              Working...
              X