Summary:
Kwin tried to not use kcrash by not linking against it.
KIO now links against KCrash as we link against KIO we indirectly link
against KCrash and thus now link and launch kcrash on wayland. This
breaks coredump.
Arguably it should be fixed in KIO, but it's non-trivial.
The advantage of this approach is it allows us to enable drkonqi for
nested mode in the future, or make a non-gui drkonqi that just saves
logs.
Test Plan: Compiles
Reviewers: #kwin, romangg
Reviewed By: #kwin, romangg
Subscribers: zzag, romangg, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D16834
Summary:
kwin_wayland now automatically chooses an appropriate backend, such as
DRM, nested wayland or nested X. If nested it will automatically go into
windowed mode regardless of whether --windowed is set and works fine.
Backend choosing logic existed duplicated in older code for kwin_wayland
--windowed, with the subtle unexpected difference that it preferred
running as an X client over running as a wayland cient if both are
present. This simplifies codes and syncs automatic resolution behaviour.
kwin_wayland --windowed with --x11-display or --wayland-display will
remain the same.
Test Plan:
Ran kwin_wayland with and without --windowed inside another wayland.
Got the same backend chosen
Tested that kwin_wayland (without --windowed) on an X machine worked just fine.
Reviewers: #kwin, mart
Reviewed By: #kwin, mart
Subscribers: mart, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D13658
Summary:
On Wayland processXcbEvents calls workspaceEvents and updateX11Time but this
is already done in the native event filter.
Test Plan:
Manual test shows Xwayland still works and constant calls to
XcbEventFilter::nativeEventFilter are processed.
Reviewers: #kwin, davidedmundson
Reviewed By: #kwin, davidedmundson
Subscribers: davidedmundson, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D14215
Summary:
KWin is not designed to run as root. It is not hardened enough and
there is a great risk that applications could attack KWin through
X11 properties, Wayland protocol requests, etc. to trigger stack or
heap overflows and execute random code. As clients connected to KWin
could be remote there is a great risk in running KWin as root. Also
clients on the same system but started as a different user could
try to gain more privs by exploiting KWin.
Furthermore KWin is designed to not run as root. It interacts with
logind to get the devices opened which would require root. Whether
KWin would work as root at all is questionable.
We cannot guarantee that running KWin as root is secure, thus this
change disallows running KWin and thus a Wayland session as root.
Reviewers: #kwin, #plasma
Subscribers: kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D13008
Summary: Now that the framebuffer device is queried, don't assume that the framebuffer device is /dev/fb0
Test Plan:
Add /dev/fb1 to seat1, start kwin with the framebuffer backend on seat1, verify with the stdio/stderr output it selected /dev/fb1.
Make sure manual device selection still works
There is an issue right now where /dev/fb1 is all black, if it is both selected manually or automatically. I think I might need see if the VSCREENINFO or FSCREENINFO framebuffer settings are changing or something (I had to do something similar to this to Weston's fbdev backend as well)
If I do find a solution, I'll send one separate from this series
Reviewers: #kwin, graesslin
Reviewed By: #kwin, graesslin
Subscribers: zzag, rkflx, graesslin, kwin, #kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D9557
Summary:
The main reason for not having it as a mandatory dependency was that BSD
doesn't support it. But as I learned recently it is available on our CI
system. So BSDs have support now.
Even more it showed that the code doesn't compile if the dependency is
missing. And there's one thing I hate: broken build configuration
options.
So let's make UDEV and libinput a required dependency and get rid of the
problems.
Test Plan: Compiles
Reviewers: #kwin, #plasma
Subscribers: plasma-devel, kwin
Tags: #plasma
Differential Revision: https://phabricator.kde.org/D10057
Once again KActivities causes a delay in the startup of KWin. This is
not the first time it happens, that has been a problem since the start
of the Wayland project.
In the past we added a compile time switch due to that and added a
runtime switch once it stoped freezing (20a9a2a247).
And now it broke again, so let's disable again.
KActivities doesn't make sense on Wayland anyway as there is no activity
support for Wayland windows. Once KActivities is able to be used without
freezing KWin on startup and being able to ensure that it won't break
again we can reenable. But I expect guarantees about that. This must be
ensured by proper autotest support to verify that in future usage of
KActivities in the display manager doesn't freeze it.
This is maintainer decision, due to that no pre-commit review.
CCMAIL: plasma-devel@kde.org
BUG: 388628
Summary:
depends from D9521
listens to switch events and updates the tablet mode status
which is exposed to dbus in the org.kde.KWin.TabletModeManager
interface
Test Plan:
as hardware support is limited, testing of clients
so far is done by the setter in the dbus property,
which should be removed from the final version.
It has been tested to successfully work on a Thinkpad.
Reviewers: #plasma, #kwin, graesslin
Reviewed By: #plasma, #kwin, graesslin
Subscribers: graesslin, davidedmundson, plasma-devel, kwin, #kwin
Tags: #plasma
Differential Revision: https://phabricator.kde.org/D9764
Summary:
Loads settings on whether QtQuick uses GL or software rendering
Aurorae needed a fix. The rest "just works" with up-to-date plasma framework.
Test Plan:
Alt tabbed
Showed an OSD
Grepped kwin's QML for any custom shaders.
(which is the only part that doesn't work here)
Reviewers: #plasma, graesslin
Reviewed By: #plasma, graesslin
Subscribers: graesslin, plasma-devel, kwin, #kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D8040
Summary:
The SCHED_RESET_ON_FORK results in threads (not only processes) to have
the scheduling policy reset to default. This means that the libinput
thread is not real time, but this was actually intended.
To solve this problem KWin does start without the RESET_ON_FORK flag
during startup. Once createInput has been called the scheduling is
adjusted again and RESET_ON_FORK is added again. This results in the
libinput thread and all threads Qt starts in between (e.g. dbus) to gain
real time policy. But it is still not leaked to other processes or to
threads in KWin which don't need it.
Other options considered: just don't use RESET_ON_FORK and instead
manually reset on fork. This would mean all threads in KWin gain real
time, but we don't need this. It's only interesting for the main
(rendering, Wayland thread) and the input thread. Also the danger to
leak into another process is too high.
Keeping the capability till the libinput thread is created and adjust
the thread itself. This option was discarded as I don't want KWin to
have any capabilities when the QApplication is started.
Test Plan: ps -eL -o class,rtprio,cmd,comm | grep kwin_wayland
Reviewers: #kwin, #plasma
Subscribers: plasma-devel, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D8065
The connect to Compositor::sceneCreated must be disconnected again,
otherwise a restart of the Compositor results in Workspace being created
again.
Thanks to our autotests for finding this problem!
The operation mode X11 and Wayland is no longer supported and probably
hasn't been supported for the last few years. But now there is the
possibility to run Wayland only, so introduce this as a dedicated new
OperationMode.
Summary:
Just continues with the steps without starting XWayland.
This does not yet fully support kwin_wayland without X11, it currently
still crashes on teardown.
Test Plan: kwin_wayland --socket=wayland-1 kwrite (note the missing --xwayland)
Reviewers: #kwin, #plasma
Subscribers: plasma-devel, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D7924
Summary:
The base idea behind this change is to keep the system responsive no
matter what other processes do. All input and rendering needs to go
through the windowing system, so keeping it responsive is important.
Currently KWin competes with all other processes for resources and this
can render the system unusable. Consider some processes running amok. In
this case the user might not be able to easily close the applications as
KWin does not get the cpu time to perform the input tasks requested by
the user.
Or in the case of playing a demanding game it is important that KWin
gets scheduled to forward the pointer input events. The user doesn't
want that the game (or another process) wins against the windowing
sytem.
The disadvantage is that KWin always wins against other processes with
real time scheduling. This could result in KWin running amok stalling
the system. On the other hand this is no change to the current situation
as if KWin runs amok the sytem is unusable.
The change uses libcap to set CAP_SYS_NICE on kwin_wayland executable.
KWin_wayland on start sets the scheduling policy to SCHED_RR with the
lowest possible priority. Thus any other SCHED_RR process will win
against KWin. So real time processes are not affected by this change!
After adjusting the scheduling (which requires CAP_SYS_NICE) KWin drops
this capability again.
Test Plan:
Verified that KWin adjusts the scheduler, that it is not passed
to child processes, that the capability gets dropped and not passed to
child processes.
Reviewers: #kwin, #plasma
Subscribers: plasma-devel, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D7757
Summary:
Provides a virtual method in Screens where backends can supply the scale
of each screen, this is then set on each output.
For the X windowed backend this value is taken from a command line
parameter.
Test Plan:
Ran windowed mode with --scale 1 and 2
then kate --platform=wayland from another screen.
On the latter case UI elements were scaled up correctly
Reviewers: #plasma
Subscribers: plasma-devel, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D3159
Summary:
Increases minimum Qt version to 5.7. This allows to drop the pre-5.7
virtual keyboard and various ifdefs for now unsupported versions.
Reviewers: #kwin, #plasma
Subscribers: plasma-devel, kwin
Tags: #kwin
Differential Revision: https://phabricator.kde.org/D4485
XWayland needs to create an eglDisplay for platform DRM. Depending on
the mesa version eglInitialize expects that the passed native display
is a Wayland display and crashes. This prevents KWin to startup.
This patch sets EGL_PLATFORM for Xwayland. This is actually not KWin's
job, but if it otherwise crashes it's better than nothing.
As Xwayland only supports DRM, it's fine to use this.
Summary:
The idea behind this mode is to support applications like spectacle
(see T4458). The calling application passes a file descriptor (created
through e.g. a pipe) and the screenshot effect writes the captured image
into that fd.
The advantage over the existing variant which writes to a file in the
/tmp directory is that this is peer-to-peer between the requesting
application and KWin. No other application can get to that image.
The change also includes setting SIGPIPE to ignore. It showed that when
the reading side already cancelled the read prior to KWin writing out
the image we get a SIGPIPE which results in application termination,
which is not what we want in case of a Wayland compositor. The sigpipe
can be ignored as Qt (and libpng) handles that error just fine at
runtime.
Reviewers: #kwin, #plasma_on_wayland
Subscribers: plasma-devel, kwin
Tags: #plasma_on_wayland, #kwin
Differential Revision: https://phabricator.kde.org/D3412
Summary:
The QStyle might interact with our internal Wayland connection. If the
cleanup happens after destroying our internal Wayland connection KWin
will crash at tear-down.
With this change the QStyle can perform cleanup already prior to the
deconstruction of the Wayland connection.
CCBUG: 372001
Reviewers: #kwin, #plasma_on_wayland
Subscribers: plasma-devel, kwin
Tags: #plasma_on_wayland, #kwin
Differential Revision: https://phabricator.kde.org/D3241
Summary:
If KWin fails to start the Wayland server due to XDG_RUNTIME_DIR not
being set, kwin_wayland should terminate with an error condition but
not crash.
This change makes sure that KWin detects that the Wayland server does
not work and terminates the startup early and ensures that it doesn't
crash while going down.
An error message is shown that we could not create the Wayland server.
Test Plan:
Test case added which verifies that WaylandServer fails to
init. Manual testing that kwin_wayland exits with error 1.
Reviewers: #kwin, #plasma_on_wayland
Subscribers: plasma-devel, kwin
Tags: #plasma_on_wayland, #kwin
Differential Revision: https://phabricator.kde.org/D2078
Summary:
We need to properly tear down the application - this can be achieved
through QCoreApplication::exit. Otherwise there is a chance that the
cleanup handling crashes.
Reviewers: #kwin, #plasma_on_wayland
Subscribers: plasma-devel, kwin
Tags: #plasma_on_wayland, #kwin
Differential Revision: https://phabricator.kde.org/D2136
Summary:
kwin_wayland disables ptrace on itself. This has the side effect of
core dumps no longer be created - which we want as DrKonqi doesn't
work for kwin_wayland.
This change introduces a dedicated signal handler for abort and
segfault. The signal handler enables ptrace again, unsets itself as
signal handler and raises the signal again, so that the proper crash,
abort handling can be performed.
Test Plan:
Added a crash, added an abort and verified that coredumpctl
shows the expected coredump.
Reviewers: #plasma_on_wayland, #kwin, bshah
Subscribers: plasma-devel, kwin
Tags: #plasma_on_wayland, #kwin
Differential Revision: https://phabricator.kde.org/D2003
Summary:
The idea is to have KWin provide a virtual keyboard. To support this
KWin uses the QT_IM_MODULE qtvirtualkeyboard and makes sure that the
QPA plugin loads it.
KWin has a new class VirtualKeyboard which acts as the focus object and
the "proxy" for input methods. The QPA plugin ensures that this is the
focusObject, so that all input method related events are sent to this
class. From there it will be possible to delegate to other applications
through the Wayland interfaces.
Reviewers: #plasma
Subscribers: plasma-devel
Tags: #plasma
Differential Revision: https://phabricator.kde.org/D1638
Summary:
Similar to[[ https://phabricator.kde.org/D1216 | D1216 ]] add procctl call to disable ptrace on FreeBSD.
We cannot do the procfs-lookup to check whether the process is already being run inside gdb -- however, on FreeBSD, we could use the P_TRACED flag of the process to figure this out:
> sys/proc.h:#define P_TRACED 0x00800 /* Debugged process being traced. */
And the code would look something similar to
```
pid_t pid = getpid();
struct procstat *prstat = procstat_open_sysctl();
struct kinfo_proc *procinfo;
unsigned int cnt;
procinfo = procstat_getprocs(prstat, KERN_PROC_PID, pid, &cnt);
long p_flags = procinfo->ki_flag;
int p_traced = p_flags & P_TRACED;
if (p_traced != P_TRACED) {
mode = PROC_TRACE_CTL_DISABLE;
procctl(P_PID, getpid(), PROC_TRACE_CTL, &mode);
}
procstat_freeprocs(prstat,procinfo);
procstat_close(prstat);
```
But as wayland is [far] in the future on FreeBSD, and that check above is a bit lengthy, I think it is enough if we add it once it is needed.
Reviewers: rakuco, graesslin
Reviewed By: graesslin
Subscribers: plasma-devel
Projects: #plasma
Differential Revision: https://phabricator.kde.org/D1425
Summary:
This change ensures that kwin_wayland does not pull in KCrash. We
don't want and need KCrash in the Wayland case. If KWin crashes the
session goes down - restarting doesn't make any sense, we need to
relogin.
Similar drkonqi just doesn't work as it doesn't have a windowing
system to connect to. After all the windowing system just crashed.
Also the AlternativeWM dialog doesn't make any sense on Wayland.
Similar thought: there is no windowing system to show this nice dialog.
Overall it's better to have system default behavior
(e.g. systemd-coredump) than using KCrash in the very special case of
kwin_wayland.
Reviewers: #plasma
Subscribers: plasma-devel
Projects: #plasma
Differential Revision: https://phabricator.kde.org/D1550
Summary:
Instead of having the Application invoke initOutputs after creating
the Screens, we can just connect to the signal emitted there.
This allows to make initOutputs a private, WaylandServer internal
method.
Reviewers: #plasma
Subscribers: plasma-devel
Projects: #plasma
Differential Revision: https://phabricator.kde.org/D1482
CI system shows that on tear down an effect might call into Workspace
after it's destroyed. Unloading all effects prior to deleting Workspace
should(TM) fix that. Unfortunately I'm not able to reproduce the problem
locally, so build.kde.org needs to verify.
So far if the Scene creation failed kwin_wayland went into a shutdown,
but didn't succeed because the thread to start Xwayland was already
running: it froze.
This change introduces a new signal in Compositor: sceneCreated. The
startup of Xwayland is bound to this signal. If it gets fired KWin can
startup Xwayland. If it does not get fired, KWin terminates correctly.
Summary:
This is the first change in a refactoring series. The aim is to:
* rename AbstractBackend to Platform
* move backends/ to plugins/platforms/
* don't bind platforms to Wayland only
* provide a platform plugin for "normal" X11
* share more code between X11 and Wayland
This change moves the platform/backend from waylandServer to Application.
The init of the plugin happens directly in the Application from the
KPluginMetaData. There is no need to externally init it and set the
parent.
WaylandServer::backend() currently just delegates to
kwinApp()->platform(), the idea is to drop this method completely.
The test infrastructure is also adjusted to this change.
Test Plan: kwin_wayland still works, all tests pass
Reviewers: #plasma, sebas
Subscribers: plasma-devel
Projects: #plasma
Differential Revision: https://phabricator.kde.org/D1331
This patch introduces additional method to Application class
that has default policy of restarting the app.
This method is overridden for wayland case disabling this step.
REVIEW: 126655
This change enables kactivities integration by default again on both
X11 and Wayland (as kactivities no longer blocks). As we have an
infrastructure to disable kactivities we can also make use of it and
offer a command line switch to disable kactivities. This might be
useful for using KWin outside of Plasma.
REVIEW: 126153
We need to unblock the signals blocked with pthread_sigmask.
This caused kdeinit to block, because it relies on SIGUSR1.
BUG: 356580
FIXED-IN: 5.5.1
REVIEW: 126361
In order to increase the security we disable ptrace on kwin_wayland.
This makes it impossible for a another process running as the same
user to attach to kwin_wayland to install a key logger. It doesn't
protect against higher privileged users, but that's no problem: they
can just read the input device file and don't need to attach to KWin
to become a key logger.
This change is highly inspired by a similar change to kscreenlocker.
A difference is that KWin checks whether we are running under a
debugger. In such a case we still want to allow ptrace.
This is similar to the applications to start. That is the value of the
command line argument is interpreted as a command to start.
The difference is that when this application exits, KWin will also quit.
The argument is so to say interpreted as a session.
Reviewed-By: Bhushan Shah
Ensures that all Wayland objects are destroyed and the cleanup handling
is performed before tearing down the Compositor. This fixes for example
a crash if a Surface with a Shadow is still around at tear down.
WaitForFinished blocks our main thread, but Xwayland wants to talk
to Wayland and blocks as well. So let's ensure events are processed
while terminating Xwayland.
Some effects do X11 calls in their cleanup code through external
libraries (e.g. KSelectionOwner). As we cannot control that we need
to ensure the effects are unloaded prior to destroying the Xwayland
connection.
We need to destroy the ClientConnections we create. Also we need
to disconnect our Xwayland error handling before destroying it, otherwise
it would trigger the abort for crashing Xwayland.
During Compositor tear down Xwayland is already destroyed. Thus it
doesn't make sense to try to delete the support properties: either
it freezes in xcb or it crashes because the connection is null.
At the same time we also ensure that the connection internally is
reset to null. Note: the one in kwinglobals.h caches and could cause
use-after-free errors. Any tear-down code must be migrated to
kwinApp()->x11Connection().
We need to destroy the compositor after Xwayland terminated and after
the internal Wayland connection is destroyed. This means when destroying
the Workspace we may no longer destroy the Compositor at the same time.
Also we need to ensure that other tear down functionality doesn't call
into the no longer existing internal client connection.
With this change kwin doesn't crash when exiting with Wayland and/or
X11 windows still open.
-use qstringliteral only when necessary (i.e. not in concat or comparison)
-use qbytearray instead of qstring when dealing with latin1 input and output (glplatform)
-use qstringref to extract numbers from strings (glplatform)
-define qt_use_qstringbuilder to optimize all string concatenations
-anidata: use ctor init lists, add windowType member initialization
REVIEW: 125933
This introduces --lockscreen option in kwin_wayland which when used will
lock screen immediately. Also dependency to newly created kscreenlocker
repo is introduced.
REVIEW: 125954
We only need to force QT_QPA_PLATFORM to be our own plugin when starting
the application. As KWin starts other processes (e.g. kcmshell5) we need
to ensure that the env variable has a sane value. No other process can
load our QPA plugin after all.
Reviewed-By: Bhushan Shah
We already unset QT_DEVICE_PIXEL_RATIO to ensure to keep out of
high dpi scaling. Now with Qt 5.6 we also need to set the attribute.
Without we crash on startup as we don't have a screen that early and
well Qt doesn't check whether it's null.
The change to reflect the available plugins in the command line options
had side effects as that made it impossible to start KWin if certain
plugins are not available.
This change ensures that we only check the option value if they are set
in the first place.
In addition the multi checks to verify whether combinations are valid
is removed as that makes in reality no sense any more (only few plugins
installed by distros) and doesn't scale any more with the increasing
numbers of backends.
A new backend which doesn't present the rendered output. It uses a
QPainter scene, renders to a QImage but doesn't present it anywhere.
Thus a real virtual backend.
By exporting the environment variable KWIN_WAYLAND_VIRTUAL_SCREENSHOTS
the backend creates a temporary dir, prints the path to std-out and
saves each rendered frame into that directory. Of course with exit it
will be deleted again.
As distributions do not ship all plugins let's try to filter the options
based on the plugins which are actually available. There is no point in
allowing e.g. framebuffer just to notice then that the plugin is not
shipped ending in confusing error for the user. So let's only add the
options which actually make sense.
Distributions do package each backend plugin in a dedicated package,
which means just because KWin got compiled with a specific backend,
doesn't mean it is also available at runtime.
In order to better support this reality this change introduces a
list-backends command line option. Each of the plugins got the json
metadata extended by the KPlugin syntax, so that we have a name and
description to print.
As we already locate all plugins anyway, the additional findPlugins for
the selected backend is changed to search the list of all plugin meta
data.
We no longer need to have the event dispatcher created before starting
the QApplication, thus we can leave it to the QPA plugin to creat it.
Also we don't need to implement our own dispatcher any more but can
use one from Qt5PlatformSupport as we link it anyways. The special
need for dispatching the WaylandServer is no longer needed as we can
explicitly dispatch it from the QPA plugin if needed.
Our QPA plugin does not require a running Wayland server at startup,
so we can move the startup after creating the application.
Nice side effect: --help works now.
The hack to perform the first eglInitialize in a thread is no longer
needed, as our internal QPA plugin is either able to share the
EGLDisplay or performs the potentially blocking eglInitialize in a
thread.
If the backend specifies
{
"input": false
}
in it's plugin metadata, it means that it does not handle input by
itself. In that case we can enforce libinput.
With this change one does no longer need to specify --libinput when
starting kwin_wayland on the drm or fbdev backend. So
kwin_wayland --xwayland
will bring up a fully functional kwin_wayland on drm platform.
Instead of setting our own filter rules we install a categories file
to allow to configure them. This seems to still have them enabled by
default, but should at least give distributions the possibility to
disable logging by default.
Looks like Qt 5.5 changed the default back to threaded and that can
result in dead locks of the compositor. So better enforce to basic
for the time being.
Another reason to have an own QPA to get rid of these "regressions".
The problem with KToolInvocation is that it creates a dead lock on
Wayland in case kdeinit is not already running. It starts kdeinit
and does a QProcess::waitForFinished and our kdeinit needs to interact
with the wayland server. So dead lock.
As KRun also calls into the dangerous code path it's no option which
leaves us with QProcess to start the processes.
A nice side-effect is that we no don't need to link KF5::Service any
more from kwin_core. Now once Plasma and Notification don't use it
any more, it will be gone completely.
kdeinit_executable doesn't make sense in the case of kwin_wayland as
start_kdeinit_wrapper will be executed after kwin_wayland thus the
framework doesn't work.
If none of the backend specific command line options are passed, we
try to autmatically determine the proper plugin.
1. if DISPLAY is set -> x11 backend
2. if WAYLAND_DISPLAY is set -> wayland backend
3. if compiled with libhybris and ANDROID_ROOT is set -> hwcomposer backend
4. if compiled with DRM -> drm backend
5. if nothing selected yet, fallback to fbdev
If an env variable is set outside of KWin (e.g. QT_QPA_PLATFORM) we
want KWin to pass the original env variable to the started process and
not pass none instead because it's one of KWin's special changed
variables.
Thus we take the environment before we adjust it and pass that to the
started processes with our own DISPLAY and WAYLAND_DISPLAY added.
So far this only allows to trigger show desktop functionality and exports
the state.
In future this should be restricted to just one dedicated desktop shell
process.
Input-method servers, like maliit, need to be known to KWin since KWin
needs to know about virtual keyboards. Virtual keyboards should be shown
as OSD layers, and they are one of the types of windows that actually
should be showable when the lock screen is active.
kwin_wayland --inputmethod /path/to/your/input-server
tries to start the input server. The input-server's window never gets
keyboard focus and is shown on top of all windows except for KWin's
internal clients.
If startup fails and there is no Workspace the Compositor was destroyed
as a child of Application with the result of being destroyed after the
Wayland server resulting in a crash.
If the Workspace gets created the Compositor will be destroyed by the
Workspace, so there's no need to destroy it early.
The QtWayland QPA plugin does not create windows with
X11BypassWindowManagerHint which means that KWin's internal windows are
not shown at all. Even worse in QtQuick it can result in a freeze due to
blocking wait for a frame rendered which can just never happen.
As windows could be created by e.g. scripts or look and feel packages we
cannot ensure that the flag will never be set in source code.
Thus we use ApplicationWayland::notify to filter for Show events and
check whether the window to be shown is bypassing window manager hits.
If it is, we unset the flag, destroy the window and show a new one.
The code didn't use QProcess as in the early days of main_wayland.cpp
Xwayland had to be started before the QGuiApplication was constructed.
As this requirement doesn't exist any more we can use QProcess and
improve the interaction by e.g. provide useful error messages if
Xwayland couldn't be started.
If we get an Expose event for any QWindow we need to ensure that all
wayland events on both client (Qt) and server (KWin) side are dispatched.
Otherwise it could happen that Qt starts rendering although the last
frame callback has not yet been emitted, causing the main gui thread to
be blocked.
QtWayland and mesa might dead lock KWin if we start rendering a QWindow
before Qt/Mesa got the last frame callback. They perform blocking wayland
event reading on the main gui thread which makes it impossible for KWin
to do the compositing and send the callback.
To workaround this problem we fake a frameRendered directly after each
damage event for a Qt internal window. Unfortunately this is not yet
completely sufficient, thus we also need to ensure that the wayland
events are processed before any events are processed which would cause
a repaint and block. Thus we first flush QtWayland's wl_display and then
our Server connection. If there were any damage events we can be sure
that the frameRendered is sent before Qt attempts to render.
QtWayland only creates popup windows if they have a parent QWindow or
if there is any window which had input. It's not enough to fake an
enter, it needs to be either a pointer button press or key press.
As KWin's useraction menu doesn't have a parent and we most likely
never send a pointer press to any QWindow it doesn't get shown. To
circumvent this we create a dummy window and fake a button press/release
on the window. After that Qt is tricked into believing there's a parent
window and shows the popup.
Faking the input is only done with at least Qt 5.5 as QtWayland crashes
on pointer event without a keymap being installed. As KWin does not yet
send keymaps we better disable the dangerous code path. With Qt 5.5 the
crash condition is fixed.
KWin used the wrong event dispatcher: QEventDispatcherUNIX insted of
QUnixEventDispatcherQPA. This caused QWindow related events never to
be send to their destination. Which is one of the reasons why KWin's
own windows are not shown at all.
As we cannot easily use QUnixEventDispatcherQPA we do the same as
that class. Inherit from QEventDispatcherUNIX and call into
QWindowSystemInterface::sendWindowSystemEvents.
This backend interacts with libhybris to create a hwcomposer which is
used for creating the egl context and surface. The initial version of
this backend is based on test_hwcomposer.cpp provided by libhybris.
Please note that using the hwcomposer backend requires a newer libepoxy,
the latest stable release is not able to bring up OpenGLES, so one needs
a master build of libepoxy.
Notes on licensing:
libhybris is Apache 2.0 licensed, which is not compatile with GPLv2.
But it is compatible with GPLv3. Thus the source files in the hwcomposer
backend are licensed GPLv3+ and not GPLv2+ as the rest of KWin. If one
uses KWin without the hwcomposer backend (which is obviously the default)
the licence doesn't change. But if the hwcomposer backend is used the
overall license of KWin changes to GPLv3+.
Each of the backends becomes a plugin. This allows kwin_wayland to load
the requested plugin and kwin itself doesn't need to link all the
libraries needed. E.g. libdrm is no longer linked if running kwin_x11.
Also this allows to create backends for the non-standard EGL platforms
(examples could be raspberrypi or Android devices).
With this change all backends need to emit the screensQueried signal
at some point. So far only x11 backend did not provide the signal,
wayland had a comparable delayed init mechanism. Now all backends use
the same mechanism.
