At the moment, our frame scheduling infrastructure is still heavily
based on Xinerama-style rendering. Specifically, we assume that painting
is driven by a single timer, etc.
This change introduces a new type - RenderLoop. Its main purpose is to
drive compositing on a specific output, or in case of X11, on the
overlay window.
With RenderLoop, compositing is synchronized to vblank events. It
exposes the last and the next estimated presentation timestamp. The
expected presentation timestamp can be used by effects to ensure that
animations are synchronized with the upcoming vblank event.
On Wayland, every outputs has its own render loop. On X11, per screen
rendering is not possible, therefore the platform exposes the render
loop for the overlay window. Ideally, the Scene has to expose the
RenderLoop, but as the first step towards better compositing scheduling
it's good as is for the time being.
The RenderLoop tries to minimize the latency by delaying compositing as
close as possible to the next vblank event. One tricky thing about it is
that if compositing is too close to the next vblank event, animations
may become a little bit choppy. However, increasing the latency reduces
the choppiness.
Given that, there is no any "silver bullet" solution for the choppiness
issue, a new option has been added in the Compositing KCM to specify the
amount of latency. By default, it's "Medium," but if a user is not
satisfied with the upstream default, they can tweak it.
The compositing timing algorithm assumes that glXSwapBuffers() and
eglSwapBuffers() block. While this was true long time ago with NVIDIA
drivers, nowadays, it's not the case. The NVIDIA driver queues
several buffers in advance and if the application runs out of them,
it will block. With Mesa driver, swapping buffer was never blocking.
This change makes the render backends swap buffers right after ending
a compositing cycle. This may potentially block, but it shouldn't be
an issue with modern drivers. In case it gets proven, we can move
glXSwapBuffers() and eglSwapBuffers() in a separate thread.
Note that this change breaks the compositing timing algorithm, but
it's already sort of broken with Mesa drivers.
One of the annoying things about EGL headers is that they include
platform headers by default, e.g. on X11, it's Xlib.h, etc.
The problem with Xlib.h is that it uses the define compiler directive to
declare constants and those constants have very generic names, e.g.
'None', which typically conflict with enums, etc.
In order to work around bad things coming from Xlib.h, we include
fixx11.h file that contains some workarounds to redefine some Xlib's
types.
There's a flag or rather two flags (EGL_NO_PLATFORM_SPECIFIC_TYPES and
EGL_NO_X11) that are cross-vendor and they can be used to prevent EGL
headers from including platform specific headers, such as Xlib.h [1]
The benefit of setting those two flags is that you can simply include
EGL/egl.h or epoxy/egl.h and the world won't explode due to Xlib.h
MESA_EGL_NO_X11_HEADERS is set to support older versions of Mesa.
[1] https://github.com/KhronosGroup/EGL-Registry/pull/111
This change moves the XRender backend to platformsupport directory,
similar to the OpenGL and the QPainter backend. This allows to put
platform-specific logic in XRenderBackend.
With 870679e46f, if the partial update
extension is unsupported, setSupportsPartialUpdate() won't be called.
The problem is that it may leave OpenGLBackend::supportsPartialUpdate()
uninitialized, which can lead to a crash if an OpenGL render backend
tries to call eglSetDamageRegionKHR() and EGL_KHR_partial_update is
unsupported.
Currently, the OpenGLBackend and the QPainterBackend have hooks to
indicate the start and the end of compositing cycle, but in both cases,
the hooks have different names. This change fixes that inconsistency.
In order to allow per screen rendering, we need the Compositor to be
able to drive rendering on each screen. Currently, it's not possible
because Scene::paint() paints all screen.
With this change, the Compositor will be able to ask the Scene to paint
only a screen with the specific id.
Currently, we use glFinish() to ensure that stream consumers don't see
corrupted or rather incomplete buffers. This is a serious issue because
glFinish() not only prevents the gpu from processing new GL commands,
but it also blocks the compositor.
This change addresses the blocking issue by using native fences. With
the proposed change, after finishing recording a frame, a fence is
inserted in the command stream. When the native fence is signaled, the
pending pipewire buffer will be enqueued.
If the EGL_ANDROID_native_fence_sync extension is not supported, we'll
fall back to using glFinish().
Besides being unused, we should avoid making OpenGL contexts current
against the EGLSurface of the first output because it's a slippery road
that may end up in leaking context-specific resources in the mid of a
compositing restart.
On Wayland, internal windows that use OpenGL are rendered into fbos,
which are later handed over to kwin. In order to achieve that, our QPA
creates OpenGL contexts that share resources with the scene's context.
The problems start when compositing has been restarted. If user changes
any compositing settings, the underlying render backend will be
reinitialized and with it, the scene's context will be destroyed. Thus,
we no longer can accept framebuffer objects from internal windows.
This change addresses the framebuffer object sharing problem by adding
a so called global share context. It persists throughout the lifetime of
kwin. It can never be made current. The scene context and all contexts
created in our QPA share resources with it.
Therefore we can destroy the scene OpenGL context without affecting
OpenGL contexts owned by internal windows, e.g. the outline visual or
tabbox.
It's worth noting that Qt provides a way to create a global share
context. But for our purposes it's not suitable since the share
context must be known when QGuiApplication attempts to instantiate a
QOpenGLContext object. At that moment, the backend is not initialized
and thus the EGLDisplay is not available yet.
BUG: 415798
Every time Platform::supportsQpaContext() is called, we go through the
list of supported extensions and perform a string comparison op. This is
not really cheap.
Commit e459c8bf54 added a sanity check to
prevent recomputing the texture matrix if the y-inverted hint hasn't been
changed, which is totally reasonable!
However, code that initializes dmabuf textures implicitly assumes that
calling setYInverted() always results in updating the matrix. But it may
be not the case if the passed value matches current isYInverted().
This change adds missing calls to force updating the texture matrix.
Note that we don't need to check the buffer size every time the dmabuf
image has been modified externally because the window pixmap is going to
be re-created if the dimensions of the attached buffer have changed.
I've seen some reports on the internet about Firefox displaying garbage
instead of videos. 99% that bug is caused by this issue. But it seems
like Firefox no longer displays corrupted videos on my machine, so it's
hard to tell.
Uses a setter and clear method pattern rather than having the code
repeated.
Instead of keeping a QPointer, now we are a QObject and we get notified
about destruction intention directly, so we can clear the pointer when
necessary.
Summary:
Notify the driver about the parts of the screen that will be repainted.
In some cases this can be benefitial. This is especially useful on lima
and panfrost devices (e.g. pinephone, pinebook, pinebook pro).
Test Plan:
Tested on a pinebook pro with a late mesa version.
Basically I implemented it, then it didn't work and I fixed it.
Maybe next step we want to look into our damage algorithm.
The main advantage of SPDX license identifiers over the traditional
license headers is that it's more difficult to overlook inappropriate
licenses for kwin, for example GPL 3. We also don't have to copy a
lot of boilerplate text.
In order to create this change, I ran licensedigger -r -c from the
toplevel source directory.
The wp_viewporter compositor extension allows clients to crop and scale
their surfaces. It can be especially useful for applications wishing to
reduce their power consumption, e.g. video players, etc.
Given that there is no any direct relationship between the surface size
and the buffer size anymore, we have to use specialized helper methods
for converting coordinates from the surface-local space to buffer pixel
space and vice versa.
The recommended way to get all EGL extension defines is to include
EGL/eglext.h. EGL/eglmesaext.h is a private header that compositors
should not use.
BUG: 422131
Summary:
Removes a few TODO items mentioning that some code was exactly the same
in a couple of different places.
drm: Remove unnecessary cast
Test Plan:
Using it right now
Will push in 2 different commits
Reviewers: #kwin, zzag
Reviewed By: #kwin, zzag
Subscribers: zzag, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D28708
Summary:
Add a small getter to query information internally if the backend supports
swap events. Defaults to true as it is the default in the GBM Wayland backend.
Test Plan: i915
Reviewers: #kwin
Subscribers: kwin
Tags: #kwin
Maniphest Tasks: T11071
Differential Revision: https://phabricator.kde.org/D25298
Summary:
We leak memory if we do not destroy the dmabuf implementation on EGL backend
going down.
Also this makes sure everything is cleaned up on shutdown.
FIXED-IN: 5.17.4
BUG: 413637
Test Plan: Compiles, settings change and shutdown ok.
Reviewers: #kwin, davidedmundson
Reviewed By: #kwin, davidedmundson
Subscribers: davidedmundson, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D25577
