Let the backends decide whether to just rebuild a derived state, or use
the NULL config to rebuild an empty logical state.
This also changes the expected screen size values of the no-outputs
test; as this case is actually handled now.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
Replace the 'scale' of an output with a vfunc on the MetaMonitorManager
class that takes a monitor and a monitor mode which calculates the
scale. On X11 this always returns 1, on KMS, the old formula is used.
On the dummy and test backends, the already configured values are
returned.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
In some cases the hardware cursor is invisible when Mutter is launched from the
TTY, due to drmModeSetCursor2 failing without a fallback being set.
This patch captures the return value of drmModeSetCursor2 and in case of an
error, enables the texture based fallback. It adds a `broken` state, that is
checked in should_have_hw_cursor() and
meta_cursor_renderer_native_realize_cursor_from_*() to avoid copying every
cursor into a gbm buffer when we know it will fail every single time.
https://bugzilla.gnome.org/show_bug.cgi?id=770020
Quick motions can come across as too fast (or slow) if it crosses outputs
with different scales. If this happens, rebuild the motion delta applying
the scale that applies to each logical monitor the pointer is crossing.
https://bugzilla.gnome.org/show_bug.cgi?id=778119
To allow for more natural pointer movements from relative pointer
devices (e.g. mouse, touchpad, tablet tool in relative mode, etc), scale
the relative motion from libinput with the scale of the monitor. In
effect, this means that the pointer movement is twice as fast (physical
movement vs numbers of pixels passed) as before, but it also means that
the same physical movement crosses the distance in a GUI no matter if
it is on a HiDPI monitor or not.
https://bugzilla.gnome.org/show_bug.cgi?id=778119
The code calculating the output scale involves calculations around pixel
and mm sizes, however we do compare post-transformation pixel sizes to
untransformed mm sizes, which breaks the DPI calculations. Fix this by
transforming back pixel sizes back to untransformed.
While we're at it, actually compare the output height to HIDPI_MIN_HEIGHT
instead of its width, it seems right according to the #define name and
comment.
https://bugzilla.gnome.org/show_bug.cgi?id=777687
The mitigation to avoid missing EDID blob was incorrect; the reason it
sometimes failed to read was a race between different applications all
trying to read the EDID at the same time. E.g. gnome-shell as GDM would
at the same time as the session gnome-shell try to read the EDID of the
same connector at the same time, triggering a race in the kernel,
making the blob reading ioctl occationally fail with ENOENT.
Remove this mitigation, as it didn't really mitigate anything; the race
could just as well happen when doing the actual read later.
https://bugzilla.gnome.org/show_bug.cgi?id=779837
When mutter is paused (i.e. not the DRM master), stop listening on
hotplug events. Instead read the current state and set modes when
resumed.
This avoids a race condition in the drm API which currently only
manages to properly deal with one application querying the EDID state
at the same time when there are multiple mutter instances running at
the same time (e.g. gnome-shell driving gdm at the same time as
gnome-shell as the session instance).
https://bugzilla.gnome.org/show_bug.cgi?id=779837
A MetaOutput is a connector, not exactly a monitor or a region on the
stage; for example tiled monitors are split up into multiple outputs,
and for what is used in input settings, that makes no sense. Change
this to use logical monitors instead of outputs.
https://bugzilla.gnome.org/show_bug.cgi?id=779745
When no output was specified, the screen limit was used to calculate the
aspect ratio. The screen limit, however, is either just an arbitrary
number if no screen limit is applicable, or a hardware graphics buffer
limit, which has nothing to do with anything actually displayed. Change
it to use the screen size instead, to get something that makes more
sense when no output is found.
https://bugzilla.gnome.org/show_bug.cgi?id=779745
Expose via a new API whether the transform on a logical monitor is
handled by the backend. This was previously only exposed only in the
native backend. This will be used to emulate not supporting transforms
in the backend in the nested backend.
https://bugzilla.gnome.org/show_bug.cgi?id=779745
Whenever an EGLOutput consumer is temporary unable to handle
eglStreamConsumerAcquire() operations (e.g. during a VT-switch),
an EGL_RESOURCE_BUSY_EXT error is generated.
This change adds the appropriate error handling to flip_egl_stream() in
order to recover from such errors.
https://bugzilla.gnome.org/show_bug.cgi?id=779112
Using ClutterInputDeviceEvdev::output-aspect-ratio. This only applies
to devices which are not calibratable, so again we need to implement
this at the toolkit level.
https://bugzilla.gnome.org/show_bug.cgi?id=774115
We couldn't properly merge output-mapping matrix and calibration into
one. Now that libinput calibration matrix is free to use, we can
actually implement tablet calibration with it.
https://bugzilla.gnome.org/show_bug.cgi?id=774115
The initial state of the hardware cursor is not known, so always force
update it the first time we update the cursor. Do this by changing the
'force' flag of update_hw_cursor() to an 'invalidated' hw cursor state.
https://bugzilla.gnome.org/show_bug.cgi?id=771056
Clutter assumed seat0 which is most usually, but not always correct.
Add an evdev-backend specific function to allow passing the seat
that will be used for ClutterDeviceManager construction, which we
already obtain in MetaLauncher.
https://bugzilla.gnome.org/show_bug.cgi?id=778092
Handle headless setup gracefully by having no logical monitors. This
commit only makes the monitor management code deal with it; other areas
may still not be able to handle it.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
The new monitor configuration system (MetaMonitorConfigManager) aims to
replace the current MetaMonitorConfig. The main difference between the
two is that MetaMonitorConfigManager works with higher level input
(MetaMonitor, MetaMonitorMode) instead of directly looking at the CRTC
and connector state. It still produces CRTC and connector configuration
later applied by the respective backends.
Other difference the new system aims to introduce is that the
configuration system doesn't manipulate the monitor manager state; that
responsibility is left for the monitor manager to handle (it only
manages configuration and creates CRTC/connector assignments, it
doesn't apply anything).
The new configuration system allows backends to not rely on deriving the
current configuration from the CRTC/connector state, as this may no longer be
possible (i.e. when using KMS and multiple framebuffers).
The MetaMonitorConfigManager system is so far disabled by default, as
it does not yet have all the features of the old system, but eventually
it will replace MetaMonitorConfig which will at that point be removed.
This will make it possible to remove old hacks introduced due to
limitations in the old system.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
Sometimes we hit a race on hot-plug where we try to read the KMS
resources and the EDID blob is not yet ready. This would normally
result in a ENOENT when retrieving the blob. Handle this by retrying
after 50 milliseconds after a hot-plug event. Do this up to 10 times,
and after that give up trying to get the EDID blob and continue with
best effort.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
The function meta_monitor_manager_read_current_config() was renamed to
meta_monitor_manager_read_current_state() as it does not read any
configuration, but reads the current state as described by the backend.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
The MetaMonitorMode referred to the mode of a CRTC, and with the future
introduction of a MetaMonitor, theh old name would be confusing.
Instead call it what it is.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
Instead of storing the logical monitors in an array and having users
either look up them in the array given an index or iterate using
indices, put it in a GList, and use GList iterators when iterating and
alternative API where array indices were previously used.
This allows for more liberty regarding the type of the logical monitor.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
Change meta_monitor_manager_get_logical_monitor_at() to use floats,
replace users of meta_monitor_manager_get_monitor_at_point() to use the
API that returns a logical monitor and remove the now unused function.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
The method used for getting the current logical monitor (the monitor
where the pointer cursor is currently at) depends on the backend type,
so move that logic to the corresponding backends.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
Let the backend initialize the cursor tracker, and change all call
sites to get the cursor tracker from the backend instead of from the
screen. It wasn't associated with the screen anyway, so the API was
missleading.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
In preparation for further refactorizations, rename the MetaMonitorInfo
struct to MetaLogicalMonitor. Eventually, part of MetaLogicalMonitor
will be split into a MetaMonitor type.
https://bugzilla.gnome.org/show_bug.cgi?id=777732
We need to do swap notifications asynchronously from flip events since
these might be processed during swap buffers if we are waiting for the
previous frame's flip to continue with the current.
This means that we might have more than one swap notification queued
to be delivered when the idle handler runs. In that case we must
deliver all notifications for which we've already seen a flip event.
Failing to do so means that if a new frame, that only swaps buffers on
such a swap notification backlogged Onscreen, is started, when later
we get its flip event, we'd notify only an old frame which would hit
this MetaStageNative's frame_cb() early exit:
if (global_frame_counter <= presented_frame_counter)
return;
and we'd never finish the new frame and thus clutter's master clock
would be waiting forever stuck.
https://bugzilla.gnome.org/show_bug.cgi?id=774557
When flush-swap-notify is already queued, we might end up trying to
requeue it, for example when handling flip callbacks inside
swap-buffers. Actually queuing it there is harmless, since old frames
will be discarded anyway.
https://bugzilla.gnome.org/show_bug.cgi?id=774923
We might still end up in swap-buffer without the previous flip callback
having been invoked. This can happen if there are two monitors, and we
manage to draw before having all monitor flip callbacks invoked.
https://bugzilla.gnome.org/show_bug.cgi?id=774923
This commit adds for a new type of buffer being attached to a Wayland
surface: buffers from an EGLStream. These buffers behave very
differently from regular Wayland buffers; instead of each buffer
reperesenting an actual frame, the same buffer is attached over and
over again, and EGL API is used to switch the content of the OpenGL
texture associated with the buffer attached. It more or less
side-tracks the Wayland buffer handling.
It is implemented by creating a MetaWaylandEglStream object, dealing
with the EGLStream state. The lifetime of the MetaWaylandEglStream is
tied to the texture object (CoglTexture), which is referenced-counted
and owned by both the actors and the MetaWaylandBuffer.
When the buffer is reattached and committed, the EGLStream is triggered
to switch the content of the associated texture to the new content.
This means that one cannot keep old texture content around without
copying, so any feature relying on that will effectively be broken.
https://bugzilla.gnome.org/show_bug.cgi?id=773629
This commit adds support for using a EGLDevice and EGLStreams for
rendering on top of KMS instead of gbm. It is disabled by default; to
enable it pass --enable-egl-device to configure.
By default gbm is first tried, and if it fails, the EGLDevice path is
tried. If both fails, mutter will terminate just as before.
https://bugzilla.gnome.org/show_bug.cgi?id=773629
There is no way to pass any backend specific parameters to a
CoglFramebuffer until after it has been allocated by
cogl_framebuffer_allocate() (since this is where the winsys/platform
fields are initialized). This can make it hard to actually allocate
anything, if the platform depends on some backend specific data.
A proper solution would be to refactor the onscreens and framebuffers to
use a GObject based type system instead of the home baked Cogl one, but
that'll be left for another day. For now, allocate in two steps, one to
allocate the backend specific parts (MetaOnscreenNative), and one to
allocate the actual onscreen framebuffer (via
meta_onscreen_native_allocate()).
So far there is nothing that forces this separation, but in the future
there will, for example EGLDevice's need to know about the CRTC in
order to create the EGLSurface.
https://bugzilla.gnome.org/show_bug.cgi?id=773629
A swap-buffers should never be issued when we are waiting for a flipped
callback, so instead of trying to handle a situation that sholud never
happen, warn instead.
https://bugzilla.gnome.org/show_bug.cgi?id=773629
When a swap failed with EACCES (possibly due to VT switching), don't
mark the framebuffer as 'in use', so that it'll be cleaned up properly
and not set as current.
https://bugzilla.gnome.org/show_bug.cgi?id=773629
For when there is no gbm available, for example when using
EGLDevice/EGLStream's, just fall back to the OpenGL texture based
cursor rendering path.
https://bugzilla.gnome.org/show_bug.cgi?id=773629
Drivers may be bad at guessing what is passed to eglGetDisplay, ending
up return non-functioning EGLDisplay's. Using eglGetPlatformDisplay
avoids this issue.
https://bugzilla.gnome.org/show_bug.cgi?id=773629
Lets use a pbuffer surface as a dummy surface instead of a gbm based
one, so that we don't need to rely on the availability of gbm to create
a dummy surface when there is no need for it.
https://bugzilla.gnome.org/show_bug.cgi?id=773629
Separate gbm initialization from general renderer initialization. Do
this even though no other initialization is done for now; later there
will will be other types of rendering mode, initialized in their own
functions.
https://bugzilla.gnome.org/show_bug.cgi?id=773629
In another step getting rid of the duplications introduced by Cogl,
use the equivalent GLib types where Cogl types previously used. While
CoglBool is not a typedef to gboolean, they are both typedefs to int,
and we already use GLib's TRUE/FALSE to set them.
https://bugzilla.gnome.org/show_bug.cgi?id=773629
Initialize the GError pointer used when creating the renderer. If an
error occurs, the error is expected to be NULL, otherwise it'll
misinterpreted as already set.
https://bugzilla.gnome.org/show_bug.cgi?id=773629
Stylus configuration (stylus buttons, pressure) was handled
at the very high level, doing the button and pressure translations
right before sending these to wayland clients.
However, it makes more sense to store these settings into the
ClutterInputDeviceTool itself, and have clutter apply the config
at the lower level so 1) the settings actually apply desktop-wide,
not just in clients and 2) X11 and wayland may share similar
configuration paths. The settings are now just applied whenever
the tool enters proximity, in reaction to
ClutterDeviceManager::tool-changed.
This commit moves all handling of these two settings to
the clutter level, and removes the wayland-specific paths
https://bugzilla.gnome.org/show_bug.cgi?id=773779
Enabling edge scrolling before disabling two finger would result in
edge scrolling not actually being enabled because two finger is still
enabled at the time and we bail out.
This patch moves this logic to common code for both the native and X
backends and fixes it by ensuring that both settings are never set at
the same time and still re-checking if edge scrolling should be
enabled after two finger scrolling gets disabled.
We also simplify the code by not checking for supported/available
settings since the underlying devices will just reject those values
and there isn't anything we can do about it here. It's the UI's job to
only show supported/available settings to users.
https://bugzilla.gnome.org/show_bug.cgi?id=771744
Checking for supported methods isn't needed since libinput will just
error out and do nothing itself if a requested method isn't supported
and, in fact, this logic was preventing the enum values 0 from being
set.
https://bugzilla.gnome.org/show_bug.cgi?id=771744
commit e2bfaf0751 does this:
g_hash_table_insert (cards,
g_udev_device_get_name (parent_device),
g_steal_pointer (&parent_device));
The problem is the g_steal_pointer call may happen before the
g_udev_device_get_name call leading to a crash.
This commit does the get_name call on an earlier line
https://bugzilla.gnome.org/show_bug.cgi?id=771442
Right now we accept any character device that matches the glob card*.
That's fine, but we can be a little more specific by checking that
the devtype is what we expect.
This commit does that.
https://bugzilla.gnome.org/show_bug.cgi?id=771442
Despite g_udev_client_new taking a list of subsystems, it doesn't
implicitly filter results to those subsystems.
This commit explicitly adds a subsystem match to make sure sound cards
don't end up in the resulting list of video cards.
https://bugzilla.gnome.org/show_bug.cgi?id=771442
Using the view's MetaMonitorInfo to find all the crtcs which should be
configured to display a given onscreen doesn't work unfortunately. The
association runs only the other way around, i.e. we need to go through
each crtc and find the ones corresponding to our monitor info.
https://bugzilla.gnome.org/show_bug.cgi?id=773115
If this isn't initialized and an idle watch gets instanced before
meta_idle_monitor_native_reset_idletime() gets called, that idle watch
would get triggered as soon as we hit the main loop.
This was causing gnome-session to go into idle mode at session start
thus making gnome-shell lock the screen.
In the past this bug was being masked by either logind emiting
session active signals or a stray input event making it through at
startup.
https://bugzilla.gnome.org/show_bug.cgi?id=772839
This isn't technically needed and, in fact, makes us default to
interlaced modes in some cases which isn't desirable.
Note that X doesn't account for these flags either for its mode
refresh rates.
https://bugzilla.gnome.org/show_bug.cgi?id=772176
As the m format specifier doesn't consume any arguments, the number
of varargs currently doesn't match the number of specifiers; the
failed transform may be relevant, so include it in the message
instead of removing the excess argument.
Some output devices only advertise their preferred mode even though
they're able to display others too. This means we can include some
common modes in each output's supported list.
This is particularly important for mirroring, since we can only mirror
outputs which are using the same resolution.
https://bugzilla.gnome.org/show_bug.cgi?id=744544
This signal allows interested parties to be notified of a new cursor
frame being painted regardless of whether it's being painted by the
backend directly or if it's a software rendered cursor frame handled
by clutter.
https://bugzilla.gnome.org/show_bug.cgi?id=749913
Switch to the output naming logic used by the X server's modesetting
driver which, in particular, uses drmModeConnector's connector_type_id
instead of connector_id.
The kernel generates new connector_id's every time there are changes
which means we can't identify the same monitor on the same connector
after an hardware hotplug. Switching to connector_type_id fixes this.
https://bugzilla.gnome.org/show_bug.cgi?id=770338
We can only honor this properly in the MUTTER_STAGE_VIEWS=1 case. When using
the legacy view, software implemented transforms are only exposed if there is
only one output, as we can only transform the entire stage there.
https://bugzilla.gnome.org/show_bug.cgi?id=745079
The texture is only created if the view is transformed at the software level,
otherwise the texture is NULL, and rendering happens on the onscreen.
https://bugzilla.gnome.org/show_bug.cgi?id=745079
The offscreen is given through the ::back-buffer property, the ClutterStageView
will set up the the CoglPipeline used to render it back to the "onscreen"
framebuffer.
The pipeline can be altered through the setup_pipeline() vfunc, so ClutterStageView
implementations can alter the default behavior of blitting from offscreen to
onscreen with no transformations.
All getters of "the framebuffer" that were expecting to get an onscreen have
been updated to call the right clutter_stage_view_get_onscreen() function.
https://bugzilla.gnome.org/show_bug.cgi?id=745079
The call to _cogl_framebuffer_winsys_update_size() results in no-op here,
as the framebuffer has already the right size when rebuilding the views.
https://bugzilla.gnome.org/show_bug.cgi?id=745079
Those will need a separate treatment from the modes that we eventually
support through "software", so split those into a separate enum so we
can can do the right thing when applying the configuration.
Also, add a helper function that returns the transform that the software
fallbacks should perform, which should be "normal" if the rotation is
already handled via hw.
The function applying the configuration has been modified to always set
a HW rotation mode (even if normal), when we come to support SW rotation
modes, we'll be relying on a normal transformation, so it will be
necessary to have mixed HW/SW managed transforms.
https://bugzilla.gnome.org/show_bug.cgi?id=745079
The scale will have been set to 1 no matter what when initializing the
MetaOutput since it at the time didn't have an CRTC assigned to it.
Now, when we assign the CRTC to the output, we need to update the scale.
https://bugzilla.gnome.org/show_bug.cgi?id=769505
Support changing the mouse and trackball acceleration profile. This
makes it possible to for example disable pointer acceleration by
choosing the 'flat' profile.
This adds an optional dependency on gudev. Gudev is used by the X11
backend to detect whether a device is a mouse or not. Without gudev
support, the accel profile settings has have effect for mouse devices.
Trackball still uses the "strstr" approach, since udev doesn't support
tagging devices as trackball devices yet.
https://bugzilla.gnome.org/show_bug.cgi?id=769179
Add support for setting edge-scrolling separately from two-finger
scrolling. We now have 2 separate boolean settings for those, with the
Mouse panel in gnome-control-center allowing to set only one of those at
a time, but nothing precludes both being set in the configuration.
We need to handle:
- two-finger-scrolling-enabled and edge-scrolling-enabled settings both
being set.
- those 2 settings being change out-of-order
- two-finger-scrolling being set on a device that doesn't support it
- edge-scrolling-enabled on a device that doesn't support it
And the combinations of one touchpad supporting just one of edge
scrolling and two-finger scrolling and another vice-versa.
https://bugzilla.gnome.org/show_bug.cgi?id=768245
