Commit Graph

29462 Commits

Author SHA1 Message Date
Jonas Ådahl
0dc6d76371 tests/kvm: Forward MUTTER_DEBUG to the virtual machine
This helps debugging the tests running using KVM.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2141>
2022-09-01 14:31:40 +00:00
Jonas Ådahl
3fe01ffc4a color-manager: Take over color device management from gsd-color
Previously, gsd-color handled adding color devices. It got information
about those via the GnomeRR API, which is part of libgnome-desktop.
libgnome-desktop itself got this information from the
org.gnome.Mutter.DisplayConfig.GetResources() D-Bus method, implemented
by mutter.

Now, mutter itself will add all the monitor color devices itself,
without having to go via gsd-color.

We sometimes need to delete colord devices synchronously, in certain
race conditions when we add and remove devices very quickly (e.g. in
tests). However, we cannot use libcolord's 'sync' API variants, as it
has a nested takes-all main loop as a way to invoke the sync call. This
effectively means we end up sometimes not return from this function in a
timely manner, causing wierd issues.

Instead, create our own sync helper, that uses a separate context that
we temporarly push as the thread-default one.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2141>
2022-09-01 14:31:40 +00:00
Jonas Ådahl
4fe743e03b dbus-runner: Add colord mocking
This will be needed for adding colord integration without breaking
testing.

The test context is altered to make sure any left over color devices are
cleaned up before starting. This means it becomes possible to run a test
case multiple times without having to restart meta-dbus-runner.py.

Note: Don't use os.getlogin() to get the current username; as that
requires a controlling terminal.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2141>
2022-09-01 14:31:40 +00:00
Jonas Ådahl
67e7140c25 monitor-manager: Move PNP lookup to MetaBackend
It's not really about monitors, even though it is used for monitors.
Lets shrink MetaMonitorManager a bit moving it to the backend.

While at it, stop leaking it too.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2141>
2022-09-01 14:31:40 +00:00
Jonas Ådahl
7ee60e1143 monitor: Don't only use 'winsys ID' to check sameness
What determines whether one MetaMonitor is the same as the other should
be whether the actual monitor is the same. The way to check this is
comparing the EDID vendor/product/serial fields. Whene these are
incomplete, fall back on the 'winsys ID'.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2141>
2022-09-01 14:31:40 +00:00
Jonas Ådahl
ccb6a7e84f monitor: Allow vendor/product/serial to return NULL
Same applies to MetaOutput. The reason for this is to make it possible
to more reliably know when there was EDID telling us about these
details. This will be used for colord integration.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2141>
2022-09-01 14:31:40 +00:00
Jonas Ådahl
f939a7ca29 tests/monitor-unit-tests: Make serials more unique
We fairly consistently had multiple monitors with the whole
vendor,product,serial tuple identical. If we start relying on making
monitors a bit more unique, e.g. for colord integration, we need to make
two monitors connected distinguishable in order for tests to properly
reflect reality and excercise the correct colord integration paths.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2141>
2022-09-01 14:31:40 +00:00
Jonas Ådahl
972e393fc0 monitor: Add API to check whether the monitor is virtual or not
As for the types of monitor, X11 and KMS are currently assumed to always be
physical, while the virtual ones are assumed to be virtual. In theory
X11 ones could be virtual, but lets not bother. KMS ones can be virtual
in the case of virtual KMS, but we typically use that for testing as if
it was physical, so lets leave it as such.

Will later be used to feed correct information to colord.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2141>
2022-09-01 14:31:40 +00:00
Jonas Ådahl
f8dbea27f8 monitor: Add API to set GAMMA LUT
It's effectively a for-each for every active CRTC of the monitor.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2141>
2022-09-01 14:31:40 +00:00
Jonas Ådahl
c6c3f340ec monitor: Add API to get GAMMA LUT size
Will be used to generate look up tables.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2141>
2022-09-01 14:31:40 +00:00
Jonas Ådahl
8446cd9f4f monitor: Add getter for EDID MD5 checksum
Will be used for colord device matching.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2141>
2022-09-01 14:31:40 +00:00
Jonas Ådahl
15bbded86b monitor-manager/native: Bail if gamma LUT size differs
Don't attempt to set a CRTC GAMMA LUT with a different size than what is
already there.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2141>
2022-09-01 14:31:40 +00:00
Jonas Ådahl
b6aea1318d util: Add 'color' debug topic
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2141>
2022-09-01 14:31:40 +00:00
Jonas Ådahl
20a91aa4e9 backend: Add color manager skeleton
Create a color manager type that eventually will be the high level
manager of color related behavior, such as ICC profiles and
color "temperature" a.k.a. night light.

For now, it's only an empty shell. It's also constructed by the actual
backend, as at a later point, the X11 and native color management
implementations will differ.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2141>
2022-09-01 14:31:40 +00:00
Jonas Ådahl
3faf1f15d7 ci: Install colord and lcms2 devel packages
Needed for colord integration and color management.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2141>
2022-09-01 14:31:40 +00:00
Jonas Ådahl
c7954be56e tests/wayland/test-driver: Unset resource user data on finalize
Client connections may linger after the test driver is teared down;
handle this gracefully by unsetting the user data on the wl_resource,
and make the resource destructor a no-op, instead of where it would
otherwise remove itself from the resource list. This fixes this crash
seen in CI:

Received signal 11 (SIGSEGV)
 #0  g_list_remove() at ../glib/glist.c:596
 #1  test_driver_destructor() at ../src/tests/meta-wayland-test-driver.c:219
 #2  destroy_resource() at ../src/wayland-server.c:730
 #3  for_each_helper() at ../src/wayland-util.c:416
 #4  wl_map_for_each() at ../src/wayland-util.c:430
 #5  wl_client_destroy() at ../src/wayland-server.c:889
 #6  wl_display_destroy_clients() at ../src/wayland-server.c:1482
 #7  meta_wayland_compositor_prepare_shutdown() at ../src/wayland/meta-wayland.c:441
 #8  meta_context_dispose() at ../src/core/meta-context.c:667
 #9  g_object_unref() at ../gobject/gobject.c:3863
 #9  g_object_unref() at ../gobject/gobject.c:3780
 #10 glib_autoptr_clear_GObject() at /usr/include/glib-2.0/gobject/gobject-autocleanups.h:29
 #10 glib_autoptr_clear_MetaContext() at ../src/meta/meta-context.h:32
 #10 glib_autoptr_cleanup_MetaContext() at ../src/meta/meta-context.h:32
 #10 main() at ../src/tests/wayland-unit-tests.c:707
 #11 __libc_start_call_main() in /usr/lib64/libc.so.6
 #12 __libc_start_main() in /usr/lib64/libc.so.6
 #13 _start() in /builds/GNOME/mutter/build/src/tests/mutter-wayland-unit

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2601>
2022-09-01 08:39:48 +00:00
Mario Limonciello
96aa0fb853 output/kms: Add more heuristics to decide when to offer fallback modes
If the panel is connected via eDP and supports more than one mode
at different resolutions don't try to add more.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2586>
2022-08-31 08:15:57 +00:00
Daniel van Vugt
bb1ffb0887 shaped-texture: Don't clear mipmaps during painting
This reverts an attempt at saving texture memory that was introduced
recently in 16fa2100. It was misguided because the same texture may be
needed in the next frame if a window has multiple previews visible on
screen at once (gnome-shell's overview). Keeping the mipmaps around
seems to reduce the peak render times of the overview by roughly 5%-10%.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2598>
2022-08-30 14:50:26 +00:00
Jiri Grönroos
c3021375e0 Update Finnish translation 2022-08-30 13:02:32 +00:00
Daniel van Vugt
8e2085dad9 backends/native: Add xlnx to the atomic deny list
Although its atomic KMS support seems to work at first, mode sets to
anything other than the Xilinx preferred max resolution of 2048x1280
would result in a hang. The xlnx kernel driver is given:
`DRM_MODE_ATOMIC_ALLOW_MODESET | DRM_MODE_PAGE_FLIP_EVENT`
and it does complete the mode set without error, but page flip events
never arrive and so you're frozen on the first frame.

Revert to legacy KMS which has no such problem with non-default modes.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2596>
2022-08-26 18:39:12 +08:00
Nart Tlisha
52b1531b95 Update Abkhazian translation 2022-08-25 11:44:21 +00:00
Carlos Garnacho
9d558e334c wayland: Unlink surface listener when freeing token
If the token ended up consumed and freed, we might leave a dangling
destroy listener after freeing the token struct.

Fixes: ed516dde89 (wayland: Add destruction listener to activation token surface)
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2594>
2022-08-25 11:15:01 +02:00
Jonas Dreßler
8ad4056aa2 clutter/click-action: Don't stop release events if gesture is inactive
I don't see how this makes sense at all, ClutterClickAction really
shouldn't mess with BUTTON_RELEASE events that are not part of a
gesture.

So propagate those events instead of stopping them.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2552>
2022-08-25 00:28:41 +00:00
Jonas Dreßler
4fce0b974b clutter: Add source field to crossing events
The source field was removed from ClutterEvent with commit
b644ea1bce because the preferred way of
getting the event actor is now to use the device/sequence actor from the
stage directly.

With crossing events it's not that easy though, as crossing events
explicitly have a source and related actor that doesn't have to be the
same actor as the device actor. Since we kept around the "related" field
there anyway, let's also introduce a "source" field in
ClutterCrossingEvent and return that actor when get_source() is called
on a crossing event.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2551>
2022-08-24 22:43:54 +00:00
Jonas Dreßler
e5784bf885 clutter/actor: Add detail to ::event signal
Just like we did with the ::captured-event signal, add detail to the
::event signal too. At the first glance this might not seem necessary
since there are individual signals like scroll-event or touch-event that
get emitted at the same time, but these don't exist for touchpad gesture
events and others.

As an easy solution for that, just make it possible to use detail on the
event signal as we did with the caputured-event signal.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2431>
2022-08-24 22:19:15 +00:00
Jonas Dreßler
84396baafb stage: Mark stage as active on wayland when it has key-focus
The "activate" and "deactivate" signals on ClutterStage are used by
Cally to track the key-focus when the user is interacting with shell UI.
key-focus only gets tracked while the stage is activated.

Wayland has no concept of the stage receiving focus or not, so right now
the activation state is bound to whether there's a focus_window in
meta_display_sync_wayland_input_focus(). Since display->focus_window is
set pretty much all the time, this effectively binds activation state to
whether the stage holds a grab or not. This is almost good enough, but
it misses cases where key-focus is on the stage without a grab, for
example when keyboard-navigating the panel after using Ctrl+Alt+Tab.

It seems to make more sense to bind the activation state to whether
key-focus is set to an actor or to NULL, so let's do that instead.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2329>
2022-08-24 10:00:22 +00:00
Carlos Garnacho
ed516dde89 wayland: Add destruction listener to activation token surface
In the timespan between an activation token being created and the
token being used by the activated application, the surface that started
the activation request may end up destroyed/disposed.

In that case, the token would be left with a stale surface pointer,
maybe causing crashes later on. Set up a destroy notification listener
so that we do know to unset the token surface if that situation arises,
this will result in Mutter not considering the token activatable, thus
maybe issuing the "Application needs attention" notification if the
activated surface did not immediately get focus. In any case this is
better than a compositor crash.

A typical situation where this may happen is "Open With..." dialogs,
since those don't live long after launching the application.

Fixes: https://gitlab.gnome.org/GNOME/mutter/-/issues/2390
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2592>
2022-08-23 22:56:42 +02:00
Carlos Garnacho
9c402bd091 wayland: Ensure to remove destroy listener for MetaWaylandBuffer
Avoid the chance of invalid memory access by leaving the wl_listener
behind when handling the destroy notification.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2588>
2022-08-23 17:06:07 +00:00
Carlos Garnacho
f4f39d3a2c wayland: Ensure to unlink destroy listeners for subsurfaces
There is some surface tracking going on here, and all notify handlers
are possibly leaving the linked wl_listener behind. Ensure it is unlinked
in all destroy notification functions.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2588>
2022-08-23 17:06:07 +00:00
Carlos Garnacho
51064c1b4e wayland: Ensure to unlink destroy listeners after destruction
This is missed in some wl_listeners added
through wl_resource_add_destroy_listener(). Ensure this is done consistently
for all notify functions.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2588>
2022-08-23 17:06:07 +00:00
Daniel van Vugt
0cfd0c3a8c monitor-store-unit-tests: Verify <maxbpc> XML is stored correctly
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2412>
2022-08-19 17:07:13 +00:00
Daniel van Vugt
26a9dd7f88 monitor-unit-tests: Add a basic set-and-get test for max_bpc
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2412>
2022-08-19 17:07:13 +00:00
Daniel van Vugt
cc203e241d monitor-test-utils: Add max_bpc field
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2412>
2022-08-19 17:07:12 +00:00
Daniel van Vugt
03336788fb monitor-config-store: Add parsing support for "maxbpc" in monitors.xml
For example:

  <monitor>
    ...
    <maxbpc>8</maxbpc>
  </monitor>

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2412>
2022-08-19 17:07:12 +00:00
Daniel van Vugt
d39137c631 monitor-manager-xrandr: Changing "max bpc" changes output assignment
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2412>
2022-08-19 17:07:12 +00:00
Daniel van Vugt
99d84ae1cc output-xrandr: Add support for the "max bpc" KMS connector property
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2412>
2022-08-19 17:07:12 +00:00
Daniel van Vugt
978cbe2da4 onscreen-native: Set max_bpc with mode sets
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2412>
2022-08-19 17:07:12 +00:00
Daniel van Vugt
f19b350c94 output-kms: Add max_bpc support
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2412>
2022-08-19 17:07:12 +00:00
Daniel van Vugt
4eef54f80a monitor-config-manager: Add a max_bpc field to MetaMonitorConfig
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2412>
2022-08-19 17:07:12 +00:00
Daniel van Vugt
a907cef905 monitor: Add function meta_monitor_get_max_bpc()
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2412>
2022-08-19 17:07:12 +00:00
Daniel van Vugt
6ff0d600c3 output: Add max_bpc support
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2412>
2022-08-19 17:07:12 +00:00
Daniel van Vugt
e0e10c29d9 monitor-manager-private: Add max_bpc to MetaOutputAssignment
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2412>
2022-08-19 17:07:12 +00:00
Daniel van Vugt
b6dbb686e2 kms-impl-device-simple: Apply META_KMS_CONNECTOR_PROP_MAX_BPC
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2412>
2022-08-19 17:07:12 +00:00
Daniel van Vugt
0cc4815837 kms-impl-device-atomic: Apply META_KMS_CONNECTOR_PROP_MAX_BPC
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2412>
2022-08-19 17:07:12 +00:00
Daniel van Vugt
2b561a369c kms-update: Add function meta_kms_update_set_max_bpc
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2412>
2022-08-19 17:07:12 +00:00
Daniel van Vugt
33ee155630 kms-connector: Add "max bpc" property
This limits the bits-per-channel colour depth of the signal that
the connector is allowed to output.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2412>
2022-08-19 17:07:12 +00:00
Daniel van Vugt
f765437410 kms-types: Add type MetaKmsRange
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2412>
2022-08-19 17:07:12 +00:00
Daniel van Vugt
f760b47a3c kms-impl-device: Remember the min/max values for a DRM_MODE_PROP_RANGE
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2412>
2022-08-19 17:07:12 +00:00
Simon McVittie
64a67aa00b backend/native: Don't warn on EACCES if headless
Since commit 1bf70334 "tests/runner: Make test runner use the headless
backend", tests are run with the native backend in headless mode, which
will attempt to open each GPU and show a warning (fatal during tests)
if it cannot.

However, in headless mode we might not be logged in on any seat (for
example we might be logged in via ssh instead), which means we might
legitimately not have permission to use any GPUs, even if they exist.
Downgrade the warning to a debug message in this case.

Resolves: https://gitlab.gnome.org/GNOME/mutter/-/issues/2381
Signed-off-by: Simon McVittie <smcv@debian.org>
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2584>
2022-08-19 16:07:23 +00:00
Daniel van Vugt
9a2c5d7d74 tools/get-state: Round refresh rates to three decimal places
Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/2465>
2022-08-19 15:28:56 +00:00