Commit Graph

114 Commits

Author SHA1 Message Date
Jonas Ådahl
bd2ca79a48 backends/x11: Only apply keymap when not running nested
https://bugzilla.gnome.org/show_bug.cgi?id=777800
2017-02-15 14:04:16 +08:00
Jonas Ådahl
6d64123849 backends/x11: Split up X11 backend into Cm and Nested
Split up the X11 backend into two parts, one for running as a
Compositing Manager, and one for running as a nested Wayland
compositor.

This commit also cleans up the compositor configuration calculation,
attempting to make it more approachable.

https://bugzilla.gnome.org/show_bug.cgi?id=777800
2017-02-15 14:04:16 +08:00
Jonas Ådahl
d188949101 backends/x11: Make the X11 backend derivable
Make the X11 backend derivable so that we can base a nested test backend
on it.

https://bugzilla.gnome.org/show_bug.cgi?id=777732
2017-01-25 16:28:53 +08:00
Jonas Ådahl
5a8509f895 Move current logical monitor retrieval to the backends
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
2017-01-25 16:28:52 +08:00
Jonas Ådahl
f5b2d79f3b backends: Use macros for declaring the backend types
https://bugzilla.gnome.org/show_bug.cgi?id=777732
2017-01-25 16:28:52 +08:00
Jonas Ådahl
d644993a74 MetaBackend: Move fields from MetaBackend into MetaBackendPrivate
Lets prepare for being able to declare the MetaBackend type using
the macro helper.

https://bugzilla.gnome.org/show_bug.cgi?id=777732
2017-01-25 16:28:52 +08:00
Olivier Fourdan
4c106a9c9b wayland: save/restore numlock state
Save the state on NumLock so that is can be (optionally) restored on
next login.

bugzilla: https://bugzilla.gnome.org/show_bug.cgi?id=757943
2016-09-09 19:07:10 +02:00
Jonas Ådahl
cc4a65fe94 x11/nested: Only paint monitor stage views when enabled
Only paint the per monitor stage views when enabled, otherwise bad
things happen.

https://bugzilla.gnome.org/show_bug.cgi?id=768976
2016-08-03 11:19:24 +08:00
Jonas Ådahl
566c28bdaf Introduce regional stage rendering
Add support for drawing a stage using multiple framebuffers each making
up one part of the stage. This works by the stage backend
(ClutterStageWindow) providing a list of views which will be for
splitting up the stage in different regions.

A view layout, for now, is a set of rectangles. The stage window (i.e.
stage "backend" will use this information when drawing a frame, using
one framebuffer for each view. The scene graph is adapted to explictly
take a view when painting the stage. It will use this view, its
assigned framebuffer and layout to offset and clip the drawing
accordingly.

This effectively removes any notion of "stage framebuffer", since each
stage now may consist of multiple framebuffers. Therefore, API
involving this has been deprecated and made no-ops; namely
clutter_stage_ensure_context(). Callers are now assumed to either
always use a framebuffer reference explicitly, or push/pop the
framebuffer of a given view where the code has not yet changed to use
the explicit-buffer-using cogl API.

Currently only the nested X11 backend supports this mode fully, and the
per view framebuffers are all offscreen. Upon frame completion, it'll
blit each view's framebuffer onto the onscreen framebuffer before
swapping.

Other backends (X11 CM and native/KMS) are adapted to manage a
full-stage view. The X11 CM backend will continue to use this method,
while the native/KMS backend will be adopted to use multiple view
drawing.

https://bugzilla.gnome.org/show_bug.cgi?id=768976
2016-07-20 14:23:48 +08:00
Jonas Ådahl
55726e787b backends: Make clutter stage resizing more explicit
The stage resizing was placed in the generic backend, which was only
run on certain configurations (when running nested or using the native
backend). This commits makes the resizing more explicit thus more
obvious.

https://bugzilla.gnome.org/show_bug.cgi?id=768976
2016-07-20 14:23:48 +08:00
Jonas Ådahl
a54042e938 Introduce MetaRenderer
MetaRenderer is meant to be the object responsible for rendering the
scene graph. It will contain the logic related to the cogl winsys
backend, the clutter backend, and the clutter stage window.

https://bugzilla.gnome.org/show_bug.cgi?id=768976
2016-07-20 14:23:48 +08:00
Jonas Ådahl
90de521799 Make mutter manage its own clutter backends
Introduce two new clutter backends: MetaClutterBackendX11 and
MetaClutterBackendNative. They are so far only wrap ClutterBackendX11
and ClutterBackendEglNative respectively, but the aim is to move things
from the original clutter backends when needed.

https://bugzilla.gnome.org/show_bug.cgi?id=768976
2016-07-20 14:23:48 +08:00
Rui Matos
7f6bcea331 compositor: Handle GL video memory purged errors
Emit a signal so that interested parties can recreate their FBOs and
queue a full scene graph redraw to ensure we don't end up showing
graphical artifacts.

This relies on the GL driver supporting the
NV_robustness_video_memory_purge extension and cogl creating a
suitable GL context. For now we only make use of it with the X backend
since the only driver with which this is useful is NVIDIA.

https://bugzilla.gnome.org/show_bug.cgi?id=739178
2016-06-27 20:23:50 +02:00
Alberts Muktupāvels
2857fdbdb8 backend-x11: Ensure the Xkb group index remains properly set
Ubuntu ships a patch in the X server that makes the group switch
keybindings only work on key release, i.e. the X server internal group
locking happens on key release which means that mutter gets the
XKB_KEY_ISO_Next_Group key press event, does its XLockGroup() call
with a new index and then, on key release, the X server moves the
index further again.

We can work around this without affecting our behavior in unpatched X
servers by doing a XLockGroup() every time we're notified of the
locked group changing if it doesn't match what we requested.

https://bugzilla.gnome.org/show_bug.cgi?id=756543
2015-10-23 14:13:26 +02:00
Rui Matos
9abc071283 backend-x11: Fallback to a default keymap if getting it from X fails
This shouldn't fail but apparently sometimes it does and in that case
having a possibly wrong idea of the keymap is still better than
crashing.

https://bugzilla.gnome.org/show_bug.cgi?id=754979
2015-09-25 19:09:48 +02:00
Florian Müllner
bc00f118f3 Avoid shadowing existing variables 2015-09-24 03:08:22 +02:00
Jonas Ådahl
8900bd2f5c backends/x11: Draw our own cursor sprite when running nested
Use a specialized cursor renderer when running as a nested Wayand
compositor. This new renderer sets an empty X11 cursor and draws the
cursor as part of the stage using the generic cursor renderer drawing
path.

https://bugzilla.gnome.org/show_bug.cgi?id=744932
2015-09-13 21:26:22 +08:00
Ting-Wei Lan
5066eaf691 build: Fix non-wayland build
This fixes build error caused by commit 614d6bd. We can simply remove
the usage of meta-wayland.c functions in non-wayland build because
META_BACKEND_X11_MODE_NESTED is only used in wayland.

https://bugzilla.gnome.org/show_bug.cgi?id=753948
2015-09-08 01:50:03 +08:00
Owen W. Taylor
614d6bd0f8 Nested X11: use KeymapNotify events to fix key state on FocusIn
If the user Alt-Tabs out of the window, we will be left thinking
the Alt key is still pressed since we don't see a release for it.

Solve this and other related issues for the nested X11 compositor
by selecting for KeymapStateMask which causes a KeymapNotify event
to be sent after each FocusIn, and when we get these events, update
the internal XKB state and send any necessary modifiers events to
clients.

https://bugzilla.gnome.org/show_bug.cgi?id=753948
2015-09-03 16:02:25 -04:00
Rui Matos
299ed424d3 compositor: Handle fences in the frontend X connection
Since mutter has two X connections and does damage handling on the
frontend while fence triggering is done on the backend, we have a race
between XDamageSubtract() and XSyncFenceTrigger() causing missed
redraws in the GL_EXT_X11_sync_object path.

If the fence trigger gets processed first by the server, any client
drawing that happens between that and the damage subtract being
processed and is completely contained in the last damage event box
that mutter got, won't be included in the current frame nor will it
cause a new damage event.

A simple fix for this would be XSync()ing on the frontend connection
after doing all the damage subtracts but that would add a round trip
on every frame again which defeats the asynchronous design of X
fences.

Instead, if we move fence handling to the frontend we automatically
get the right ordering between damage subtracts and fence triggers.

https://bugzilla.gnome.org/show_bug.cgi?id=728464
2015-08-13 14:14:46 +02:00
Florian Müllner
cba2ab445e backend-x11: Don't assert on something that might happen
While we shouldn't normally receive crossing events for any windows
except the stage when running nested, we do in case we hold a pointer
grab - just ignore those events instead of taking down the user's
session.
2015-08-10 17:00:57 +02:00
Rui Matos
39763d4add compositor: Add support for GL_EXT_x11_sync_object
If GL advertises this extension we'll use it to synchronize X with GL
rendering instead of relying on the XSync() behavior with open source
drivers.

Some driver bugs were uncovered while working on this so if we have
had to reboot the ring a few times, something is probably wrong and
we're likely to just make things worse by continuing to try.  Let's
err on the side of caution, disable ourselves and fallback to the
XSync() path in the compositor.

https://bugzilla.gnome.org/show_bug.cgi?id=728464
2015-08-07 17:27:42 +02:00
Jasper St. Pierre
2cbaa6660c backend-x11: Fix our awful hack for crossing events, too
For enter / leave events, which we use in the UI code, we need to make
sure that these coordinates are root-relative as well, otherwise the
cursor when entering frames might be incorrect.
2015-06-24 10:35:38 -07:00
Carlos Garnacho
82a7060cdb x11: Fix touch events on nested wayland
If we're running as a nested compositor, we must not attempt to
passive grab on the root window, and we should be setting the
touch event mask on the stage window.

https://bugzilla.gnome.org/show_bug.cgi?id=751036
2015-06-16 20:39:46 +02:00
Jasper St. Pierre
6aead0c67c backend-x11: Use a mode switch to determine whether we're nested or not
The "meta_is_wayland_compositor" checks were getting a bit difficult to read.
2015-06-16 08:13:13 -07:00
Carlos Garnacho
af9d8f1931 backends/x11: Call meta_backend_update_last_device() from XI_DeviceChanged
On X11, calling this function on meta_display_handle_events() will not catch
mouse events happening over clients, so poke directly in the backend for
XI_DeviceChanged events, which mutter will get on device switches.

The code has been slightly refactored so we deal with XIEvents at a single
handle_input_event() function.

https://bugzilla.gnome.org/show_bug.cgi?id=712775
2015-03-13 21:01:11 +01:00
Jasper St. Pierre
7e7c8ecbd4 backend-x11: Also spoof Enter/Leave notifies as well
So that we track when the user enters/exits the frame window
appropriately. This fixes a rogue cursor appearing when the window
doesn't define one.
2015-01-19 21:56:32 -08:00
Jasper St. Pierre
b9fb4a5887 backend: Use a GHashTable for device monitors
The array code has been tricky to maintain and leaky. Let's just use a
GHashTable to simplify our lives.
2014-12-31 08:53:57 -08:00
Rui Matos
478b75e803 backend-x11: Re-upload keymap when new keyboard devices are added
The X server applies a default keymap to hotplugged keyboard
devices. To enforce our current settings we must re-upload the keymap
when a new keyboard shows up.

Note that setting the VCK keymap causes the server to propagate it
to all slave keyboard devices.

https://bugzilla.gnome.org/show_bug.cgi?id=737673
2014-10-02 19:14:03 +02:00
Owen W. Taylor
602dd7fdf2 MetaBackendX11: Don't XResizeWindow behind clutter's back
In the case of a nested Wayland compositor inside an X session,
Clutter is managing the toplevel window size, so don't call
XResizeWindow on it - this will confuse Clutter and get the size
and the hints out of sync on the toplevel window.

https://bugzilla.gnome.org/show_bug.cgi?id=736279
2014-09-08 14:40:50 -04:00
Jasper St. Pierre
39f65f9f86 backend: Add a keymap-changed signal 2014-08-14 17:36:29 -04:00
Jasper St. Pierre
bda2d6d1ac backend-x11: Don't leak the keymap
meta_backend_get_keymap is supposed to return a static keymap, not a new
one every time. Cache it internally. We don't update it when the keymap
changes on the server, but we'll do this soon.
2014-08-14 17:35:49 -04:00
Jasper St. Pierre
467465c99c backend: Create the stage in the backend, rather than the compositor
This allows creating the stage much earlier than it otherwise would have
been. Our initialization sequence has always been a bit haphazard, with
first the MetaBackend created, then the MetaDisplay, and inside of that,
the MetaScreen and MetaCompositor.

Refactor this out so that the MetaBackend creates the Clutter
stage. Besides the clarity of early initialization, we now have much
easier access to the stage, allowing us to use it for things such as
key focus and beyond.
2014-08-13 20:08:46 -04:00
Rui Matos
101b215d6b backends: Add methods to handle keymaps
These methods allow us to set and get xkbcommon keymaps as well as
locking a specific layout in a layout group.

With this, we introduce dependencies on xkeyboard-config, xkbfile,
xkbcommon-x11 and a libX11 new enough to have xcb support.

https://bugzilla.gnome.org/show_bug.cgi?id=734301
2014-08-07 11:24:24 +02:00
Jasper St. Pierre
4b5a503cee backend-x11: Make sure to not crash during initialization
Sometimes we can get a host event without having the display up and
running yet. Just don't pass it to the compositor in that case, since it
won't be possible for it to have any event that matters.
2014-07-27 16:14:46 +02:00
Adel Gadllah
6ee5a1437c x11: Move event filter to backend connection 2014-07-19 18:51:26 +02:00
Adel Gadllah
029673d0ee Revert "backend-x11: Take the touch grab on the stage window"
This reverts commit 3b85e4b2b9.

This breaks touch support; reverting would break wayland
(is what this patch tried to fix; we should find a better solution
that works on both).
2014-07-17 17:18:08 +02:00
Jasper St. Pierre
3b85e4b2b9 backend-x11: Take the touch grab on the stage window
So we don't get assert fails when touching the screen on mutter nested.
2014-07-14 17:24:31 -04:00
Jasper St. Pierre
55354f5c38 Take the touch grab in MetaBackendX11
This means that it won't trigger when we're running as a Wayland
compositor, which will cause a few warnings and possibly more
brokenness.
2014-07-14 09:41:10 -04:00
Carlos Garnacho
704cae1de3 backend: Use the most recent event time on replayed touch events
When a touch sequence is passively grabbed and later rejected, events
will be replayed on the next client in propagation order, although those
events (either transformed to pointer events or not) will contain the
original timestamps, this will make grabs fail with InvalidTime if triggered
from the replayed ButtonPress/TouchBegin handler.

In order to work around this, store the most recent event time (presumably
gotten from the XI_TouchEnd caused by the passive grab being rejected), and
use that time on the events being replayed afterwards and grabs, so we don't
possibly fail with InvalidTime if those events result in a compositor grab.
2014-07-08 11:21:29 -04:00
Carlos Garnacho
faf55c4627 backend-x11: Ensure touch event coordinates are translated to the stage's
Now that those events are going to be received, coordinates translation
should also happen on those.
2014-07-07 19:17:03 -04:00
Carlos Garnacho
f351c5d304 backend-x11: Remove pointless goto
The function is not as complex as needing that, plus it always jumped
anytime bypass_clutter is TRUE, so make all conditional code depend on
that.
2014-07-07 19:15:48 -04:00
Jasper St. Pierre
63bca17ab6 backend-x11: Only spoof the event when we're passing it to Clutter
There's no need to do so otherwise.
2014-07-07 19:13:31 -04:00
Jasper St. Pierre
b704659899 backend: Add an XIWarpPointer wrapper
This makes Alt+F7 / Alt+F8 work respectively under X11 nested mode.
For the native backend implementation, we'll need a special Clutter
function, so don't implement that for now.
2014-05-27 14:28:35 -04:00
Jasper St. Pierre
480a853263 x11: Add simple cursor renderer 2014-05-17 13:27:16 -04:00
Jasper St. Pierre
c8f3820956 backend-x11: Always pass None for the cursor
This means we fall back to what's defined on the root window, which
will give us our properly set cursor.
2014-05-13 16:20:03 -04:00
Jasper St. Pierre
9607513e62 backend-x11: Fix cursor position under nested 2014-04-28 13:29:45 -04:00
Jasper St. Pierre
18730f2a5e backend-x11: Take grabs on the stage window
This means that grabs will be reported with respect to the
stage window, and that our event coordinates will be more
correct under nested.
2014-04-28 13:29:31 -04:00
Jasper St. Pierre
c5d4f4a245 backend-x11: Split out event translation to another method
Keep the indentation from getting too hairy.
2014-04-24 12:12:54 -04:00
Jasper St. Pierre
a027937ccc backend-x11: Set the proper coordinates on the spoofed device event
When we click on a window with a passive grab, then the event_x
and event_y will be relative to that window, instead of relative to
the stage, which means that picking will be wrong.

Forcibly using root_x / root_y breaks nested mode. Nested mode is
a testing mode that should be replaced by a DRI3-enabled Xephyr,
though. It's getting too hairy to support properly.
2014-04-24 12:12:54 -04:00
Jasper St. Pierre
01a773ed5a backend-x11: Don't use clutter_stage_get_default
It turns out Clutter doesn't actually set the stage as the default
when creating a new one, so this creates a new stage that then dies.
2014-04-23 16:01:02 -04:00
Jasper St. Pierre
97074ccdad backend-x11: Fix leak of the event data 2014-04-23 15:43:34 -04:00
Jasper St. Pierre
6111bb9dec backend-x11: Adapt a consistent naming scheme for the event struct 2014-04-23 15:43:34 -04:00
Jasper St. Pierre
e9eb3c32a9 compositor: Move event spoofing code to MetaBackendX11
This is now where we handle device events.
2014-04-23 14:23:13 -04:00
Jasper St. Pierre
fdd43fc2d9 backend-x11: Add KeyPress / KeyRelease to mask 2014-04-23 14:23:12 -04:00
Jasper St. Pierre
1dc8a7eca3 monitor: Remove handle_xevent as a generic method 2014-04-23 12:15:12 -04:00
Jasper St. Pierre
861cfc23df x11: Make sure to handle RANDR events from the right connection 2014-04-23 12:15:12 -04:00
Jasper St. Pierre
86368e7e07 backend: Add grab_device / ungrab_device
This makes sure that we take active grabs on the correct connection.
Passive grabs are still broken.
2014-04-23 12:15:12 -04:00
Jasper St. Pierre
6c743dad88 xrandr: Rewrite to be based on the X11 display connection 2014-04-23 10:24:26 -04:00
Jasper St. Pierre
28666da6a1 backend-x11: Simplify XSync initialization handling a bit 2014-04-22 14:17:08 -04:00
Jasper St. Pierre
3d091e514d backend: Poll events from the host X11 server ourselves
I was accidentally pulling events from the Xwayland server under
nested for the idle monitor, which is wrong. Whoops.
2014-04-22 10:26:03 -04:00
Jasper St. Pierre
c44b1d730d backends: Move MonitorManager creation to MetaBackend as well 2014-04-21 20:25:21 -04:00
Jasper St. Pierre
48dc544bef backends: Move most of the code from meta_backend_init into the subclasses 2014-04-21 20:25:21 -04:00
Jasper St. Pierre
31d744195d backends: Build MetaBackend subclasses for each backend 2014-04-21 20:25:21 -04:00