Commit Graph

7042 Commits

Author SHA1 Message Date
Sebastian Keller
71101cab82 clutter/cogl/stage: Remove duplicate region union
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1089
2020-02-26 07:26:10 +00:00
Sebastian Keller
c4b28b0939 clutter/cogl/stage: Don't leak FB damage region on redraw
offset_scale_and_clamp_region() creates a new region resulting in
view_damage which at this point is the only thing left pointing to what
originally was fb_damage getting overwritten and being leaked.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1089
2020-02-26 07:26:10 +00:00
Jonas Ådahl
159c83b914 clutter: Move redraw clip management to views
The stage window handled the redraw clip in a global manner; this would
interfere if we want to paint views individually as it'd mean
intersecting views (i.e. mirrored monitors) would loose the redraw clip
once the first view was painted. It also is awkward to have a global
state for something that is built up before redrawing, and only really
valid during paint, due to buffer damage history.

This commits removes all redraw clip management from the stage window,
moving it all into the stage views. When a redraw clip is added to the
stage, every affected view will get the same redraw clip added to it,
and eventually when painted, the stage window (ClutterStageCogl) will
retrieve the redraw clip for each view as it repaints them.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1042
2020-02-25 19:18:54 +01:00
Jonas Ådahl
b68c630329 clutter/cogl/stage: Use g_clear_pointer() to clean up regions
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1042
2020-02-25 19:18:54 +01:00
Jonas Ådahl
fe1ccea1e1 clutter: Pass redraw clip via paint context
Instead of users fetching it via `clutter_stage_get_redraw_clip()`, pass
it via the paint context. This is helpful as it is only valid during a
paint, making it more obvious that it needs to be handled differently
when there is no redraw clip (i.e. we're painting off-screen).

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1042
2020-02-25 19:18:54 +01:00
Jonas Ådahl
c483b52d24 clutter/stage: Pass redraw clip instead of extents when painting view
That's the struct we have ready, the callee can just call
cairo_region_get_extents() if it only cares about the extents rectangle.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1042
2020-02-25 19:18:54 +01:00
Jonas Ådahl
f3dcba27f5 clutter/cogl/stage: Simplify region transform code
Add a helper that scales and clamps a region, aimed to be used when
transforming between framebuffer coordinate space and view coordinate
spaces.

This helps readability by moving out the verbose for loops that deals
with the individual rects of a region to the helper, making the logic
where it's used much simpler.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1042
2020-02-25 19:18:54 +01:00
Jonas Ådahl
86ccc28413 clutter/cogl/stage: Rename have_clip to is_full_redraw
The 'have_clip' variable has repeatedly confused me to meaning that
there is a clip. What it actually means is that the effective clip
covers the whole view; the 'redraw_clip == NULL' meaning full redraw is
an important implementation detail for the context, and makes the
intention of the variable unclear; especially since we will after a
couple of blocks will *always* have a clip, just that it covers the
whole view.

Rename the variable to 'is_full_redraw' and negate the meaning, aiming
to make things a lot more clear.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1042
2020-02-25 18:39:51 +01:00
Jonas Ådahl
0d039c3ba3 clutter/view: Don't take paint rect after paint
It was unused, and will be become unused by the primary paint function
as well, so go ahead and remove it here first.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1042
2020-02-25 18:39:51 +01:00
Jonas Ådahl
adcbc2aa5f clutter/cogl/stage: Use fb size for for fallback fb clip region
When calculating the fallback framebuffer clip region, which should be
the region in framebuffer coordinates, we didn't scale the view layout
with the view framebuffer scale, meaning for any other scale than 1,
we'd draw a too small region of the view. Fix this by just using the
size of the framebuffer directly, avoiding any scale dependent
calculation all together.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1042
2020-02-25 18:39:51 +01:00
Jonas Ådahl
860906246e clutter/cogl/stage: Simplify swap_event boolean logic
We'll expect a swap event if any of the view paints resulted in a swap;
make the logic dealing with this clearer by making changing the less
vilible '|| swap_event' postfix with a up front '|=' operator.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1042
2020-02-25 18:39:51 +01:00
Jonas Ådahl
37d7df612b clutter/stage/cogl: Name variable to make coordinate space obvious
The ambiguous "clip" was renamed to "fb_clip_region", as it was called
at the call site. This should make it more clear that the clip is in
buffer coordinate space.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1042
2020-02-25 18:39:51 +01:00
Jonas Ådahl
0baf4578c8 clutter/stage/cogl: Remove some leftover whitespace
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1042
2020-02-25 18:39:51 +01:00
Jonas Ådahl
1c98f01a65 renderer-native: Draw stage separately per CRTC
Prior to this commit the stage was drawn separately for each logical
monitor. This allowed to draw different parts of the stage with
different transformations, e.g. with a different viewport to implement
HiDPI support.

Go even further and have one view per CRTC. This causes the stage to
e.g. draw two mirrored monitors twice, instead of using the same
framebuffer on both. This enables us to do two things: one is to support
tiled monitors and monitor mirroring using the EGLStreams backend; the
other is that it'll enable us to tie rendering directly to the CRTC it
will render for. It is also a requirement for rendering being affected
by CRTC state, such as gamma.

It'll be possible to still inhibit re-drawing of the same content
twice, but it should be implemented differently, so that it will still
be possible to implement features requiring the CRTC split.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1042
2020-02-25 18:39:51 +01:00
Jonas Ådahl
292d2754ae clutter/stage: Remove stray newline
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1042
2020-02-25 18:39:51 +01:00
Christian Hergert
d327cedb83 clutter: avoid redundant _clutter_paint_node_init_types()
This only needs to be initialized once but is in the hot path of creating
new paint nodes (for which we create many). Instead, do this as part of
the clutter_init() workflow to keep it out of the hot path.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1087
2020-02-25 01:14:09 +00:00
Christian Hergert
0c55e87d8f cogl: reduce temporary allocations calculating redraw regions
When calculating regions, a lot of temporary allocations are created. For
the array of rects (which is often a short number of them) we can use
stack allocations up to 1 page (256 cairo_rectangle_int_t). For building
a region of rectangles, cairo and pixman are much faster if you have all
of the rectangles up front or else it mallocs quite a bit of temporary
memory.

If we re-use the cairo_rectangle_int_t array we've already allocated (and
preferably on the stack), we can delay the creation of regions until after
the tight loop.

Additionally, it requires fewer allocations to union two cairo_region_t
than to incrementally union the rectangles into the region.

Before (percentages are of total number of allocations)

     TOTAL    FUNCTION
[ 100.00%]    [Everything]
[ 100.00%]      [gnome-shell --wayland --display-server]
[  99.67%]        _start
[  99.67%]          __libc_start_main
[  99.67%]            main
[  98.60%]              meta_run
[  96.90%]                g_main_loop_run
[  96.90%]                  g_main_context_iterate.isra.0
[  96.90%]                    g_main_context_dispatch
[  90.27%]                      clutter_clock_dispatch
[  86.54%]                        _clutter_stage_do_update
[  85.00%]                          clutter_stage_cogl_redraw
[  84.98%]                            clutter_stage_cogl_redraw_view
[  81.09%]                              cairo_region_union_rectangle

After (overhead has much dropped)

     TOTAL    FUNCTION
[ 100.00%]    [Everything]
[  99.80%]      [gnome-shell --wayland --display-server]
[  99.48%]        _start
[  99.48%]          __libc_start_main
[  99.48%]            main
[  92.37%]              meta_run
[  81.49%]                g_main_loop_run
[  81.49%]                  g_main_context_iterate.isra.0
[  81.43%]                    g_main_context_dispatch
[  39.40%]                      clutter_clock_dispatch
[  26.93%]                        _clutter_stage_do_update
[  25.80%]                          clutter_stage_cogl_redraw
[  25.60%]                            clutter_stage_cogl_redraw_view

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1071
2020-02-24 22:44:35 +00:00
Christian Hergert
81d11ac0a8 clutter: avoid g_signal_emit_by_name() from ClutterActor
g_signal_emit_by_name() is used to emit signals on ClutterContainer when
actors are removed or added. It happens to do various interface lookups
which are a bit unneccessary and can allocate memory.

Simply using emission wrappers makes all of that go away.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1083
2020-02-24 22:36:27 +00:00
Jonas Dreßler
a9b1134dfb clutter/seat: Add API to inhibit unfocus of the cursor surface/actor
Add API to ClutterSeat that allows inhibiting the unsetting of the
pointer focus surface. This can be useful for drawing custom cursor
textures like the magnifier of gnome-shell does.

In the future this API should also control unsetting of Clutters
focus-actor, not just the focus surface, that's not really needed right
now since we never unset the focus-actor anyway.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1077
2020-02-24 09:52:06 +00:00
Christian Hergert
4c421959dc clutter: fix hole in ClutterPaintNode
Fixing the missalignment takes the structure from 80 bytes down to 72.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1081
2020-02-21 22:36:31 +00:00
Jonas Dreßler
d7c7311ceb clutter/seat: Add gtkDoc comments for touch-mode property
It's not that easy to find out when the touch-mode is enabled without
reading the code, so document that.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1075
2020-02-21 15:24:58 +00:00
Daniel van Vugt
282b09c17e clutter/actor: Add CLUTTER_OFFSCREEN_REDIRECT_ON_IDLE
Which offscreens actor rendering only in cases where it hasn't changed for
2 frames or more. This avoids the performance penalty of offscreening an
actor whose content is trying to animate at full frame rate. It will
switch automatically.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1069
2020-02-21 18:07:46 +08:00
Christian Hergert
d122b66abc clutter/actor: avoid transform node for identity matrix
If the transform matrix is an identity, then positioning wont change and
we can avoid creating the transform node altogether. This is based on
a similar find in GTK today while reducing temporary allocations.

This cuts the number of transforms created in clutter_actor_paint() by
about half under light testing of GNOME Shell from 6.8% to 2.4% of
allocations.

Before:

    ALLOCATED      TOTAL    FUNCTION
[   20.4 MiB] [  21.20%]    clutter_actor_paint
[   11.0 MiB] [  11.45%]      clutter_paint_node_paint
[    6.6 MiB] [   6.84%]      clutter_transform_node_new
[    2.5 MiB] [   2.61%]      clutter_actor_node_new

After:

    ALLOCATED      TOTAL    FUNCTION
[   33.4 MiB] [  24.12%]    clutter_actor_paint
[   26.2 MiB] [  18.91%]      clutter_paint_node_paint
[    3.4 MiB] [   2.43%]      clutter_actor_node_new
[    3.3 MiB] [   2.41%]      clutter_transform_node_new

Allocation amounts will have differed due to different amounts of running
time, but the % of allocations has now dropped below
clutter_actor_node_new() which should be expected.

https://gitlab.gnome.org/GNOME/mutter/issues/1056
2020-02-20 06:41:34 +00:00
Daniel van Vugt
b70611e3cf clutter/enums: Correct docs about ClutterOffscreenRedirect
The default value is zero, not the flag the docs say.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1068
2020-02-19 17:36:29 +00:00
Florian Müllner
35fe6a40ed clutter: Make clutter_contraint_update_preferred_size() public
This is necessary for gobject-introspection to correctly pick up
the annotations for the corresponding vfunc.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1067
2020-02-19 11:57:05 +00:00
Emmanuele Bassi
4f8e518d42 bind-constraint: Update the preferred size
Bind the preferred size of an actor using a BindConstraint to the
preferred size of the source of the constraint, depending on the
coordinate of the constraint.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1070
2020-02-19 08:26:28 -03:00
Adam Jackson
5c704e3f81 cogl: Remove no-op cogl_shader_compile and cogl_shader_get_info_log
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1024
2020-02-12 21:55:44 +00:00
Carlos Garnacho
dcba7c5044 clutter: Add ClutterSeat::touch-mode boolean property
And the corresponding getter. This property returns FALSE by default
and must be overridden by subclasses. This will allow gnome-shell to
hook up specific behavior that should not happen on mouse+keyboard
setups.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1044
2020-02-10 22:04:53 +00:00
Sebastian Keller
2212a44cc2 clutter/actor: Fix GValue leak when overwriting transition values
When the 'initial'/'final' values of an existing transition were being
overwritten, the GValues used in _clutter_actor_create_transition()
were not being unset.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1033
2020-02-01 15:00:15 +00:00
Sebastian Keller
09af21cd0b clutter: Fix input device node path leak
https://gitlab.gnome.org/GNOME/mutter/merge_requests/1030
2020-01-30 22:28:56 +01:00
Daniel van Vugt
2731be6929 clutter/actor: Use the new shallow relayout API
If an actor sets flag `CLUTTER_ACTOR_NO_LAYOUT` then that means it
is (or should be) unaffected by `queue_relayout` calls in its children.
So we can avoid propagating `queue_relayout` all the way up to the stage
and avoid a full stage relayout each time.

But those children whose parent has `CLUTTER_ACTOR_NO_LAYOUT` still need
to be allocated at some point. So we do it at the same point where it
happened before. Only we now queue a *shallow* relayout so the `allocate`
run on the next frame doesn't need to descend the whole actor tree anymore.
Only a subtree and hopefully very small.

For free-floating and top-level actors this provides a measurable
performance benefit. According to Google Profiler, calls to
`_clutter_stage_maybe_relayout` are now so cheap that they no longer show
up in performance profiles.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/575
2020-01-30 18:31:24 +00:00
Daniel van Vugt
5257c6ecc2 clutter/stage: Add an API for shallow relayouts
Where "shallow" means you can specify any actor instead of every relayout
having to start from the stage.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/575
2020-01-30 18:31:24 +00:00
Carlos Garnacho
1e9682b417 backends: Move warp_pointer() to ClutterSeat
The onscreen pointer sprite is a per-seat element, so it makes sense
to move pointer warping over there too.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/852
2020-01-30 18:12:58 +01:00
Carlos Garnacho
d3160d095d clutter: Drop ClutterDeviceManager
This is mostly replaced by ClutterSeat, which offers a per-seat instead
of a global device abstraction.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/852
2020-01-30 18:11:23 +01:00
Carlos Garnacho
f8fa86f9db clutter: Replace ClutterDeviceManager usage in favor of ClutterSeat
https://gitlab.gnome.org/GNOME/mutter/merge_requests/852
2020-01-30 18:11:23 +01:00
Carlos Garnacho
835f12043e clutter: Drop select_stage_events() device manager vfunc
This is specific to X11, so handle it within the X11 backend.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/852
2020-01-30 18:02:34 +01:00
Carlos Garnacho
7afbc01fe8 clutter: Move motion compression handling to ClutterSeat
Another responsibility taken away from ClutterDeviceManager

https://gitlab.gnome.org/GNOME/mutter/merge_requests/852
2020-01-30 18:02:31 +01:00
Carlos Garnacho
1c689b83f9 clutter: Move virtual device management to ClutterSeat
A11y in general may be considered a per-seat feature.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/852
2020-01-30 18:02:14 +01:00
Carlos Garnacho
8d0527f674 clutter: Move pointer a11y from ClutterDeviceManager to ClutterSeat
A11y in general may be considered a per-seat feature.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/852
2020-01-30 18:01:53 +01:00
Carlos Garnacho
bc7ba8c875 clutter: Move keyboard a11y from ClutterDeviceManager to ClutterSeat
A11y in general may be considered a per-seat feature.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/852
2020-01-30 18:01:50 +01:00
Carlos Garnacho
34ce39f8dc clutter: Move platform event data management to ClutterSeat
https://gitlab.gnome.org/GNOME/mutter/merge_requests/852
2020-01-30 17:59:47 +01:00
Carlos Garnacho
9be4f98c73 clutter: Move keymap from ClutterBackend to ClutterSeat
Keymaps are a per-seat feature, so move it there.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/852
2020-01-30 17:59:47 +01:00
Carlos Garnacho
1fc021c3c6 clutter: Move bell_notify() from ClutterBackend to ClutterSeat
Bells can be considered to be a per-seat feature, so move it there.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/852
2020-01-30 17:59:46 +01:00
Carlos Garnacho
9aba37f70f clutter: Move keymap direction from ClutterBackend to ClutterKeymap
This makes more sense to have in the ClutterKeymap, since we have it.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/852
2020-01-30 17:59:46 +01:00
Carlos Garnacho
467e94f858 clutter: Add ClutterSeat object
This object will take most responsibilities away from ClutterDeviceManager,
with the peculiarity that it'll express the per-seat device tree.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/852
2020-01-30 16:28:51 +01:00
Olivier Fourdan
8b3bc1f97d clutter: Export clutter_input_device_update()
The private function `_clutter_input_device_update()` is not currently
exported.

This function calls `_clutter_input_device_set_actor()` which updates
the `ClutterActor` under the pointer, so making that function available
outside of Clutter will allow to make sure the pointer device actor is
updated prior to do picking.

Also, now that the functions is exported to the upper layers, drop the
underscore suffix from the function name.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1026
2020-01-30 09:37:58 +01:00
Jonas Dreßler
f2f89d9b46 clutter/actor: Respect CONTENT_SIZE request mode in get_preferred_* APIs
Right now the CONTENT_SIZE request mode for a ClutterActor is only
respected by `clutter_actor_get_preferred_size()`, but not by
`get_preferred_width()` and `get_preferred_height()`. Those simply try
to ask the layout manager and will return [0, 0] for actors without
children. So be consistent and also return the content size in those two
functions.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1019
2020-01-27 14:57:13 +00:00
Daniel van Vugt
d28a640960 backend-eglnative: Search for the xsettings schema recursively
As recommended by the docs for `g_settings_schema_source_get_default`:

> The returned source may actually consist of multiple schema sources
> from different directories, depending on which directories were given
> in XDG_DATA_DIRS and GSETTINGS_SCHEMA_DIR. For this reason, all lookups
> performed against the default source should probably be done recursively.

Now it's actually found and works again, including subpixel font smoothing.

Closes: https://gitlab.gnome.org/GNOME/gnome-shell/issues/1447

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1017
2020-01-24 17:38:06 +00:00
Jonas Dreßler
b3fbc26340 clutter/actor: Allow setting width and height properties to -1.0
ClutterActors width and height can be reset to automatically use the
preferred (calculated) value by setting the width or height to -1, so
far this only works by setting it using `clutter_actor_set_width()` or
`clutter_actor_set_height()`, make sure it can also be done using the
"width" and "height" GObject properties.

https://gitlab.gnome.org/GNOME/mutter/merge_requests/1018
2020-01-24 17:27:31 +00:00
Sebastian Keller
65401ed3f9 clutter/cogl: Plug region leak in clutter_stage_cogl_redraw_view
Fixes https://gitlab.gnome.org/GNOME/mutter/issues/1000
2020-01-16 20:57:20 +00:00