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
If the surfaceless context extension is unsupported by the underlying
platform, the QPA will use the EGLSurface of the first output to make
OpenGL contexts current.
If an internal window attempts to make an OpenGL context current while
compositing is being restarted, for example it's typically the case with
the composited outline visual, QPA will either try to make the context
current with a no longer valid EGLSurface for the first output or will
crash during the call to Platform::supportsSurfacelessContext(). The
latter needs more explanation. After the compositingToggled() signal has
been emitted, there is no scene and supportsSurfacelessContext() doesn't
handle this case.
In either case, we could return EGL_NO_SURFACE if compositing is being
restarted, but if the underlying platform doesn't support the surfaceless
context extension, then the composited outline will not be able to
delete used textures, framebuffer objects, etc.
This change addresses that problem by making sure that every platform
window has a pbuffer allocated in case the surfaceless context extension
is unsupported.
SurfaceInterface::inputIsInfinite() has been dropped. If the surface has
no any input region specified, SurfaceInterface::input() will return a
region that corresponds to the rect of the surface (0, 0, width, height).
While the new design is more robust, for example it's no longer possible
to forget to check SurfaceInterface::inputIsInfinite(), it has shown some
issues in the input stack of kwin.
Currently, acceptsInput() will return false if you attempt to click the
server-side decoration for a surface whose input region is not empty.
Therefore, it's possible for an application to set an input region with
a width and a height of 1. If user doesn't know about KSysGuard or the
possibility of closing apps via the task manager, they won't be able to
close such an application.
Another issue is that if an application has specified an empty input
region on purpose, user will be still able click it. With the new
behavior of SurfaceInterface::input(), this is no longer an issue and it
is handled properly by kwin.
In order to determine if a window accepts input at specified position,
we need to check if the decorated window and the server-side decoration
contain the given point.
While we have a way to determine if a point falls inside the input shape
of the decorated window, there is no any way to do the same for the deco.
Currently, flip output transformations in the software fallback code
path are equivalent to normal rotate output transformations.
This change implements flip output transformations according to the
wl_output spec.
Currently, when the DRM platform uses cursor planes, the cursor on
a rotated output may be cropped because the math behind the current
cursor transform matrix is off.
In order to fix the cropping issue, this change replaces the current
cursor transform matrix with the core part of the surface-to-buffer
matrix, which was written against the wl_output spec.
BUG: 427605
CCBUG: 427060
Currently, every time compositing is restarted, both the gbm and the egl
streams backend will re-obtain the EGLDisplay object.
This is wrong because the core assumption is that the EGL display doesn't
change once it has been obtained.
Currently, Qt clients send two maximize requests separated by the
initial commit. From spec's perspective, this is totally fine, the
client should receive two configure events with "maximized" state.
But because changeMaximize() in XdgToplevelClient and setMaximized()
operate on two different maximize modes, the second maximize request
will trick kwin into thinking that the client should be restored.
If Xwayland has crashed, the Workspace will block stacking order updates
and start destroying all X11 clients.
Once stacking order updates are unblocked, the Workspace will mark the X
stacking order as dirty and create a new Xcb::Tree object.
We don't want to create that Xcb::Tree object because accessing it
after the XCB connection has been shut down will lead to a crash.
BUG: 427688
FIXED-IN: 5.20.1
Placement::placeTransient() checks the frame geometry right after
setting it. That is a problem because geometry updates for
XdgPopupClient are made in async fashion. We need to block geometry
updates in order to ensure that window placement code sees correct
geometry.
On some devices, the GPU nodes are also added as /dev/dri/cardX, they
are not useful for KMS purposes and does not have display resources.
If we encounter such cards, then skip them.
This check is completely wrong for mobile GPUs where GPU are internal
and are not attached through PCI subsystem,
P: /devices/platform/display-engine/drm/card1
N: dri/card1
L: 0
S: dri/by-path/platform-display-engine-card
E: DEVPATH=/devices/platform/display-engine/drm/card1
E: DEVNAME=/dev/dri/card1
E: DEVTYPE=drm_minor
E: MAJOR=226
E: MINOR=1
E: SUBSYSTEM=drm
E: USEC_INITIALIZED=4239383
E: ID_PATH=platform-display-engine
E: ID_PATH_TAG=platform-display-engine
E: ID_FOR_SEAT=drm-platform-display-engine
E: DEVLINKS=/dev/dri/by-path/platform-display-engine-card
E: TAGS=:master-of-seat:seat:uaccess:
For example, on A64 platform, the KMS capable card is card1 which is not
attached with PCI.
This makes our QPlatformOpenGLContext private subclass simpler.
As a slightly unrelated change, this patch also fixes a bug where our
platform opengl context may return a wrong surface format if surfaceless
contexts are unsupported.
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.
When we destroy all internal clients and override-redirect windows, we
assume that corresponding lists will be implicitly detached.
However, in some cases, that might be not the case. For example, if the
list is not shared, neither begin() nor end() will detach. Therefore, it
is possible to hit invalidated iterators if the list is modified inside
the loop.
This change prevents hitting invalidated iterators by making explicit
list copies.
CCBUG: 427373
- Do not set emit the enabledChanged of the text-input-v2 when it
requests us to show input panel, it is already enabled when you get
this request
- If we get the request to show input panel, reset the inputContext,
there's two reason we can get this event: 1) when keyboard focus is
activated or 2) when keyboard focus is changed to text field in same
surface. In both cases, information in input method context is not
valid anymore and should be reset.
- Also instead of ignoring the update_state events from client, adopt
input method accordingly.
