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.
EGL for X and EGL for Wayland backends are quite different. The main
motivation behind this change is to prepare the EGL backends for
monitoring vblank events. Things work quite differently depending on
if the EGL backend renders onto a toplevel window or overlay window.
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.
Since the Screens class is a convenience wrapper around AbstractOutput
objects that come from the Platform, it should not be platform-specific.
By dropping createScreens(), output-related code becomes simpler.
These signals can be useful if you want to know what output exactly has
been disabled or enabled.
The outputEnabled signal is emitted after the outputAdded signal, and
the outputDisabled signal is emitted before the outputRemoved signal.
At the moment, the gbm_device for the primary device is destroyed before
the EGLDisplay is destroyed. This results in a crash in Mesa.
In order to fix the crash, this change ensures that the EGLDisplay is
destroyed before the gbm device.
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.
At the moment, despite the protocol supporting it, we were not feeding
the EDIDs. KScreen was falling back to the output name so it didn't fail
horribly but it's still a good idea to provide all the data.
If a cursor animation is driven purely by frame callbacks and kwin
uses hardware cursors, the cpu usage may spike to 100%.
This change addresses that issue by sending frame callbacks after a
compositing cycle has been performed.
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.
Summary:
windowIcon.pixmap().bits() creates and deletes the QPixmap object.
Which means data is dangling when we call m_winInfo->setIcon()
Surprisingly harmless in real life.
Test Plan: Reran kwin_wayland on X11 with valigrind
Reviewers: #kwin, apol
Reviewed By: apol
Subscribers: zzag, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D28667
Summary:
AbstractOutput::name() behaviour is changed so that it matches the X11
behaviour, showing an identifier like "HDMI-0".
XdgOutput.name is set to this name.
XdgOutput.description is currently set to the manufacturer name and
model, but it's not exposed to Qt so we probably don't care too much.
This should fix plasmashell changing applets when switching between X11
and wayland.
Test Plan:
Relevant unit test
I still need to run it on my laptop.
Reviewers: #kwin, #plasma, zzag
Reviewed By: #kwin, #plasma, zzag
Subscribers: kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D28230
Summary:
As is KWin only had 1 Cursor which was a singleton. This made it impossible for
us to properly implement the tablet (as in drawing tablets) support and show where
we're drawing.
This patch makes it possible to have different Cursors in KWin, it makes all the
current code still follow the mouse but the tablet can still render a cursor.
Test Plan: Tests pass, been using it and works as well as before but with beautiful tablet cursors.
Reviewers: #kwin, cblack, davidedmundson
Reviewed By: #kwin, cblack, davidedmundson
Subscribers: davidedmundson, cblack, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D28155
We have lots of inconsistency at the moment in CMakeLists.txt files. Most
of it is due to kwin being a very old project. This change hopefully fixes
all of it.
Summary:
We repeat quite a lot of code that finds an output by xcb_window_t and
translates global X11 screen coordinates to output coordinates.
Test Plan: Compiles.
Reviewers: #kwin, gladhorn
Reviewed By: #kwin, gladhorn
Subscribers: gladhorn, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D23947
Summary:
Make it more explicit what the relation is between Wayland and XDG objects
existing and enablement:
The ouput is enabled if and only if Wayland and XDG output objects exist.
We can simplify the code by replacing checks on the outputs with checking
the current enablement value.
Test Plan: Wayland nested and virtual backends.
Reviewers: #kwin
Subscribers: kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D23553
Summary:
With f013a4369c, cd6b69a4d2, d960be4b3f and edb200f6bc all possible
backends of a Wayland session have been ported to using the internal functions
of AbstractWaylandOutput's for managing outputs.
This removes the alternative code path used before in these backends, what
simplifies the code and is also a prerequisite to removing the Screens global
in the future.
Reviewers: #kwin, apol, zzag
Reviewed By: #kwin, apol, zzag
Subscribers: apol, kwin
Tags: #kwin
Maniphest Tasks: T11459, T11098
Differential Revision: https://phabricator.kde.org/D23485
Summary:
Get the physical size directly from the always available output device
interface instead of saving an additional copy in the abstract wayland
output class.
There is some ambiguity with orientation and naming that needs to be
cleaned up when output orientation is reworked.
Test Plan: Nested Wayland, Drm, virtual backends tested.
Reviewers: #kwin, davidedmundson
Reviewed By: #kwin, davidedmundson
Subscribers: kwin
Tags: #kwin
Maniphest Tasks: T11459
Differential Revision: https://phabricator.kde.org/D23496
Summary:
Get the pixel size directly from the always available output device
interface instead of saving an additional copy in the backends.
Test Plan: Nested Wayland, Drm, virtual backends tested.
Reviewers: #kwin, davidedmundson
Reviewed By: #kwin, davidedmundson
Subscribers: kwin
Tags: #kwin
Maniphest Tasks: T11459
Differential Revision: https://phabricator.kde.org/D23495
Summary:
So far we were following a bit unique and rare doxygen comment style:
/**
* Contents of the comment.
**/
Doxygen comments with this style look balanced and neat, but many people
that contribute to KWin don't follow this style. Instead, they prefer
more traditional doxygen comment style, i.e.
/**
* Contents of the comment.
*/
Reviewing such changes has been a bit frustrating for me (so selfish!)
and for other contributors.
This change switches doxygen comment style in KWin to a more traditional
style. The main reason for doing this is to make code review process easier
for new contributors as well us.
Reviewers: #kwin, davidedmundson
Reviewed By: #kwin, davidedmundson
Subscribers: kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D22812
Summary:
Currently code base of kwin can be viewed as two pieces. One is very
ancient, and the other one is more modern, which uses new C++ features.
The main problem with the ancient code is that it was written before
C++11 era. So, no override or final keywords, lambdas, etc.
Quite recently, KDE compiler settings were changed to show a warning if
a virtual method has missing override keyword. As you might have already
guessed, this fired back at us because of that ancient code. We had
about 500 new compiler warnings.
A "solution" was proposed to that problem - disable -Wno-suggest-override
and the other similar warning for clang. It's hard to call a solution
because those warnings are disabled not only for the old code, but also
for new. This is not what we want!
The main argument for not actually fixing the problem was that git
history will be screwed as well because of human factor. While good git
history is a very important thing, we should not go crazy about it and
block every change that somehow alters git history. git blame allows to
specify starting revision for a reason.
The other argument (human factor) can be easily solved by using tools
such as clang-tidy. clang-tidy is a clang-based linter for C++. It can
be used for various things, e.g. fixing coding style(e.g. add missing
braces to if statements, readability-braces-around-statements check),
or in our case add missing override keywords.
Test Plan: Compiles.
Reviewers: #kwin, davidedmundson
Reviewed By: #kwin, davidedmundson
Subscribers: davidedmundson, apol, romangg, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D22371
Summary:
QImage::byteCount() was deprecated in Qt 5.10. It is advised to use
QImage::sizeInBytes() method instead.
Reviewers: #kwin, apol
Reviewed By: apol
Subscribers: kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D22355
Summary:
Represent outputs in the X11 session via AbstractOutput. For that we
move all Wayland specific parts of AbstractOutput into a new subclass
AbstractWaylandOutput and let the outputs of our Wayland backends inherit
from there.
This should allow us to get rid of the Screens class later on.
Test Plan: Manually in X session.
Reviewers: #kwin, zzag, davidedmundson
Reviewed By: #kwin, zzag, davidedmundson
Subscribers: ngraham, zzag, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D19208
Summary:
To homogenize our backends and as another step to remove the Screens class
use the AbstractOutput class in the windowed X11 backend.
Test Plan: Manually in X session.
Reviewers: #kwin, zzag
Reviewed By: #kwin, zzag
Subscribers: davidedmundson, zzag, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D19207
Summary:
So far KWin didn't send axis_source, axis_discrete, and axis_stop. Even
though most of those events are optional, clients need them to work as
expected. For example, one needs axis_source and axis_stop to implement
kinetic scrolling; Xwayland needs axis_discrete to prevent multiple
scroll events when the compositor sends axis deltas greater than 10, etc.
BUG: 404152
FIXED-IN: 5.17.0
Test Plan:
* Content of a webpage in Firefox is moved by one line per each mouse
wheel "click";
* Scrolled gedit using 2 fingers on GNOME Shell, sway, and KDE Plasma;
in all three cases wayland debug looked the same (except diagonal scroll
motions).
Reviewers: #kwin, davidedmundson
Reviewed By: #kwin, davidedmundson
Subscribers: davidedmundson, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D19000
Summary:
On Wayland we cannot switch from OpenGL to QPainter compositor as this
would break any running OpenGL application. KWin registers it's
EGLDisplay to Wayland and without OpenGL this doesn't make sense any
more. We are not able to render OpenGL buffers in the QPainter
compositor.
While it's theoretically possible to switch from QPainter to OpenGL it
doesn't make any sense for the same reason. Any running OpenGL
application would be using llvmpipe and could not be switched to proper
OpenGL.
This change stores the selected compositing type in Platform and the
implementations can use it to restrict the supported compositors. On X11
we don't need this, all other Platforms implement the restriction. Thus
it's no longer possible to switch the backends at runtime.
Test Plan:
Adjusted tests run, no runtime test as gui doesn't support
switching to QPainter anyway.
Reviewers: kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D19084
Summary:
This change inits XInput extension, listens for touch events and
forwards them to our platform API. Thus touch events are forwarded on a
nested wayland session on X11.
Please note that I only tested this change on Xwayland.
Test Plan: Run nested kwin_wayland with two outputs and looked into debug console
Reviewers: #kwin
Subscribers: kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D17369
