In order for a surface to be used as a cursor plane framebuffer, it
appears that requiring the buffer to be linear is sufficient.
GBM_BO_USE_SCANOUT is added in case GBM_BO_USE_LINEAR isn't sufficient
on untested hardware.
Fixes#1323
Removed wlr_drm_plane.cursor_bo as it does not serve any purpose
anymore.
Relevant analysis (taken from the PR description):
While trying to implement a fix for #1323, I found that when exporting
the rendered surface into a DMA-BUF and reimporting it with
`GBM_BO_USE_CURSOR`, the resulting object does not appear to be valid.
After some digging (turning on drm-kms debugging and switching to legacy
mode), I managed to extract the following error: ```
[drm:__setplane_check.isra.1 [drm]] Invalid pixel format AR24
little-endian (0x34325241), modifier 0x100000000000001 ``` The format
itself refers to ARGB8888 which is the same format as
`renderer->gbm_format` used in master to create the cursor bo. However,
using `gbm_bo_create` with `GBM_BO_USE_CURSOR` results in a modifier of
0. A modifier of zero represents a linear buffer while the modifier of
the surface that is rendered to is `I915_FORMAT_MOD_X_TILED` (see
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tree/include/uapi/drm/drm_fourcc.h?h=v4.20.6#n263).
In order to fix this mismatch in modifier, I added the
`GBM_BO_USE_LINEAR` to the render surface and everything started to work
just fine. I wondered however, whether the export and import is really
necessary. I then decided to test if the back buffer of the render
surface works as well, and at least on my hardware (Intel HD 530 and
Intel UHD 620) it does. This is the patch in this PR and this requires
no exporting and importing.
I have to note that I cheated in order to import DMA_BUFs into a cursor
bo when doing the first tests, since on import the Intel drivers check
that the cursor is 64x64. This is strange since cursor sizes other than
64x64 have been around for quite some time now
(https://lists.freedesktop.org/archives/mesa-commit/2014-June/050268.html).
Removing this check made everything work fine. I later (while writing
this PR) found out that `__DRI_IMAGE_USE_CURSOR` (to which
`GBM_BO_USE_CURSOR` translates) has been deprecated in mesa
(https://gitlab.freedesktop.org/mesa/mesa/blob/master/include/GL/internal/dri_interface.h#L1296),
which makes me wonder what the usecase of `GBM_BO_USE_CURSOR` is. The
reason we never encountered this is that when specifying
`GBM_BO_USE_WRITE`, a dumb buffer is created trough DRM and the usage
flag never reaches the Intel driver directly. The relevant code is in
https://gitlab.freedesktop.org/mesa/mesa/blob/master/src/gbm/backends/dri/gbm_dri.c#L1011-1089
. From this it seems that as long as the size, format and modifiers are
right, any surface can be used as a cursor.
We create the EGL config with GBM_FORMAT_ARGB8888, but then initialize GBM BOs
with GBM_FORMAT_XRGB8888. This mismatch confuses Mesa.
Instead, we can always use GBM_FORMAT_ARGB8888, and use DRM_FORMAT_XRGB8888
when calling drmModeAddFB2.
Fixes https://github.com/swaywm/wlroots/issues/1438
Frame events group logically connected pointer events. It makes sense to make
the backend responsible for sending frame events, since once the events are
split (ie. once the frame events are stripped) it's not easy to figure out
which events belongs to which frame again.
This is also how Weston handles frame events.
Fixes https://github.com/swaywm/wlroots/issues/1468
This makes compositors able to block and/or customize set_selection requests
coming from clients. For instance, it's possible for a compositor to disable
rich selection content (by removing all MIME types except text/plain). This
commit implements the design proposed in [1].
Two new events are added to wlr_seat: request_set_selection and
request_set_primary_selection. Compositors need to listen to these events and
either destroy the source or effectively set the selection.
Fixes https://github.com/swaywm/wlroots/issues/1138
[1]: https://github.com/swaywm/wlroots/issues/1367#issuecomment-442403454
The noop backend is similar to headless, but it doesn't contain a
renderer. It can be used as a place to stash views for when there's no
physical outputs connected.
In particular, modified public creator and destructor function names,
added a display destroy listener, safely extract user data from
resources, send correct time (in usecs) in rootston, etc.
This is a common interface that can be used for all primary selection
protocols, as discussed in [1]. A new function wlr_seat_set_primary_selection
is added to set the primary selection for all protocols.
The seat now owns again the source, and resets the selection to NULL when
destroyed.
[1]: https://github.com/swaywm/wlroots/issues/1367#issuecomment-442403454
This commits completely refactors wlr_gtk_primary_selection. The goal is to
remove gtk-primary-selection state from the seat and better handle inert
resources where it makes sense.
wlr_seat_client.primary_selection_devices has been removed and replaced by
wlr_gtk_primary_selection_device. This allows us to make offers inert when the
current selection is replaced.
wlr_seat_set_primary_selection has been removed because it relied on wlr_seat
instead of wlr_gtk_primary_selection_device_manager. A new function,
wlr_gtk_primary_selection_device_manager_set_selection (candidate for the
longest function name in wlroots) has been added. It doesn't take a serial
anymore as serial checking only makes sense for set_selection requests coming
from Wayland clients (serial checking is now done in the Wayland interface
implementation).
Since wlr_gtk_primary_selection_device_manager is now required to set the
selection, a new function wlr_xwayland_set_gtk_primary_selection_device_manager
(candidate number two for longest function name) has been added.
Devices are now made inert when the seat goes away.
Future work includes removing the last primary selection bits from the seat,
mainly wlr_seat.primary_selection_source and wlr_seat.events.primary_selection,
replacing those with new fields in wlr_gtk_primary_selection_device. Or maybe
we could keep those in the seat and replace them with a re-usable interface
(for future zwp_primary_selection_v1 support). We need to think how we'll sync
these three protocols (GTK, X11 and wayland-protocols).
See https://github.com/swaywm/wlroots/issues/1388
wlr_subsurface_from_wlr_surface can return NULL if the wl_surface is still
alive and if the wl_subsurface has been destroyed. Make sure we check for NULL.
Fixes https://github.com/swaywm/sway/issues/3195
This commit makes it possible for a single client to have multiple data devices
for the same seat. This fixes issues with Firefox.
This mainly removes wlr_data_source.offer. We make sure we create one data
offer per device. We now make the offer inert when the source is destroyed.
Fixes the second half of https://github.com/swaywm/wlroots/issues/1041
The read format is dependent on the output, so we first need to make it
current. This fixes a race condition in wlr-screencopy-v1 where a dmabuf
client would cause EGL_NO_SURFACE to be bound at the time when
screencopy needs to query for the preferred format, causing GL errors.
When a client was creating multiple data devices for the same seat, we were
only creating one resource. This is a protocol error.
Instead, create one offer per data device.
This commit also makes offers inert when their source is destroyed.
Fixes part of https://github.com/swaywm/wlroots/issues/1041
Supersedes https://github.com/swaywm/wlroots/pull/1113
The renderer redesign is going to need the render fd before the backend
is fully started, so we have to move the wl registry code to when the
backend is created instead of when it is started.
We also need to stash the wl_keyboard and emit it to library users
later, once they've added their listeners and started the backend.
We were assuming GL_BGRA_EXT was always supported.
We now check that it's supported for rendering. We fail if it isn't because
this format is specified as "always supported" by the Wayland protocol.
We also check if it's supported for reading pixels. A new preferred_read_format
function returns the preferred format that can be used to read pixels. This is
used by the screencopy protocol.
There was a missing copy_drm_surface_mgpu call in drm_connector_schedule_frame
so we asked for a pageflip with an unknown BO, resulting in ENOENT.
Additionally, this commit makes schedule_frame return a bool indicating
failures. This allows schedule_frame_handle_idle_timer to only set
frame_pending to true if a frame has been successfully scheduled. Thus, if a
pageflip fails, rendering won't be blocked forever anymore.
In case a pageflip is already pending, true is returned because a frame has
already been scheduled and will be sent sometime soon.
shm_open is a POSIX function creating an in-memory file. Using it simplifies
the code and removes the dependency on XDG_RUNTIME_DIR. The only downside is
that we need to generate a random name for the shm file.
If a client uses an older version of the dmabuf protocol, use the
`formats` event instead of `modifiers` (since that didn't exist in older
versions).
With a bit of necessary guessing, support dmabuf importing even when
EGL_EXT_image_dma_buf_import_modifiers isn't present instead of
failing up front.
The compositor acts as a relay between applications using the text-input protocol and input methods using the input-method protocol.
This change implements the basic but useful support for input-method, leaving out grabs as well as popups.
Since the source doesn't always come from a client, this field
doesn't make sense. It is replaced by a new "finalized" field in
wlr_client_data_source. This is used to make sure set_actions is
not sent after start_drag has been sent.
A check in data_offer_choose_action has been removed: if an offer
has been sent then start_drag has been called, no need to check.
I also wanted to add a check for wl_data_source.offer, but it
turns out (1) this isn't in the spec (2) it breaks GTK+.
This is some preliminary work for Firefox on Wayland compatibility.
This desynchronizes our rendering loop with the vblank cycle.
In case a compositor doesn't swap buffers but schedules a frame,
emitting a frame event immediately enters a busy-loop.
Instead, ask the backend to send a frame when appropriate. On
Wayland we can just register a frame callback on our surface. On
DRM we can do a no-op pageflip.
Fixes#617Fixesswaywm/sway#2748
This calculates and returns the effective damage of the surface in
surface coordinates, including the client damage (in buffer
coordinates), and damage induced by resize or move events.
When a pageflip is pending, we'll get a DRM event for the connector
in the future. We don't want to free the connector immediately
otherwise we'll use-after-free in the pageflip handler.
This commit adds a new state, "DISAPPEARED". This asks the pageflip
handler to destroy the output after it's done pageflipping.
Originally I asumed tilt_x and tilt_y are very unlikely to change
independent, I was proven wrong.
And while investigating Krita not using the Erasor tool, I found a bug,
which is unrelated though.
* Rename the constraint_create signal to new_constraint for
consistency
* Move the constraint_destroy signal to the constraint itself
* Use rotate_child_position instead of duplicating logic
* Fix inert constraint resource handling
* Style fixes
Sessions can now be retrieved from a backend in a more general manner.
Multi-backend gets back its `session` field that contains the session
if one was created, removing the interfacing from multi backend with the
drm backend directly. This adds the possibility to use sessions even
without the drm backend.
It additionally fixes the bug that 2 session objects got created when
WLR_BACKENDS were set to "libinput,drm".
To allow vt switching without drm backend (and drm fd) on logind, start
listening to PropertiesChanged signals from dbus and parse the session
"Active" property when no master fd was created (this does not change
current drm backend behaviour in any way).
To prevent wl_keyboard keymap being written to by clients, use a unique
file descriptor for each wl_keyboard resource.
Reference: weston, commit 76829fc4eaea329d2a525c3978271e13bd76c078
This commit allows outputs that need a CRTC to steal it from
user-disabled outputs. Note that in the case there are enough
CRTCs, disabled outputs don't loose it (so there's no modeset
and plane initialization needed after DPMS). CRTC allocation
still prefers to keep the old configuration, even if that means
allocating an extra CRTC to a disabled output.
CRTC reallocation now happen when enabling/disabling an output as
well as when trying to modeset. When enabling an output without a
CRTC, we realloc to try to steal a CRTC from a disabled output
(that doesn't really need the CRTC). When disabling an output, we
try to give our CRTC to an output that needs one. Modesetting is
similar to enabling.
A new DRM connector field has been added: `desired_enabled`.
Outputs without CRTCs get automatically disabled. This field keeps
track of the state desired by the user, allowing to automatically
re-enable outputs when a CRTC becomes free.
This required some changes to the allocation algorithm. Previously,
the algorithm tried to keep the previous configuration even if a
new configuration with a better score was possible (it only changed
configuration when the old one didn't work anymore). This is now
changed and the old configuration (still preferred) is only
retained without considering new possibilities when it's perfect
(all outputs have CRTCs).
User-disabled outputs now have `possible_crtcs` set to 0, meaning
they can only retain a previous CRTC (not acquire a new one). The
allocation algorithm has been updated to do not bump the score
when assigning a CRTC to a disabled output.
sx, sy used to store the buffer offset of the drag surface which was
then be added (by rootston) to the drag icon position.
Buffer offsets are handled already in surface_intersect_output
(output.c) so they were added twice for dnd surfaces.
This commit handles better situations in which the number of
connected outputs is greater than the number of available CRTCs.
It'll enable as many outputs as possible, and transfer CRTCs to
outputs that need one on unplug.
This changes CRTC and plane reallocation to happen after scanning
DRM connectors instead of on modeset.
This cleanups CRTCs and planes on unplug to allow them to be
re-used for other outputs.
On modeset, if an output doesn't have a CRTC, the desired mode is
saved and used later when the output gains a CRTC.
Future work includes giving priority to enabled outputs over
disabled ones for CRTC allocation. This requires the compositor to
know about all outputs (even outputs without CRTCs) to properly
modeset outputs enabled in the compositor config file and disable
outputs disabled in the config file.
This is so we can potentially add comments to it, and so if a user looks
at the installed header, they can see the /* #undef WLR_HAS_FEATURE */
line to see every option, even if not available.
Implement the tablet-v2 tablet tool's implicit grab semantics for
buttons and tip.
This avoids losing focus (to other [sub]surfaces) when a button is held,
or the tip is down.
This should help when the device is used close to a surface's border and
would otherwise have to be very precise.
This introduces -DWLR_USE_UNSTABLE and adds information regarding the
stability status to all headers. I started with a conservative set of
headers to mark as stable:
- types/wlr_matrix.h
- util/edges.h
- util/log.h
- util/region.h
- xcursor.h
Whenever a new surface is created, we have to update the cursor focus,
even if there's no input event. So, we generate one motion event, and
reuse the code to update the proper cursor focus. We need to do this
for all surface roles - toplevels, popups, subsurfaces.
Fixes#1162
153f37bdf5 (#1145) removed the
wlr_xwayland_is_unamanged function while fixing OR, because it was
belieived that it's supposed to work around the broken OR handling.
This was a misunderstanding. is_unmanaged is (while sort of a hack)
intended to work around inherent differences between "real" X sessions
and our Xwayland/wayland situation.
The main reason it exists is to support applications like rofi and dzen,
while not handing focus to other OR windows (which should *not* be
required).
Traditionally, these applications just grabbed input from X and didn't
need to be focused by any logic in the WM. Which of course doesn't work
in wayland compositors. So we have to give them focus in some way.
Giving *every* OR window focus, breaks other applications that don't
expect focus to change.
A testcase that was pointed out to me where wlr_xwayland_is_unamanged was
breaking things is https://github.com/swaywm/sway/issues/2128 (syncplay,
gitk, gitgui)
Supposedly it broke using keyboard to navigate the menus.
I can't reproduce this with this patch. The popups can be navigated as
long as the parent has focus.
Implement the basic logic for tablet-v2 tablet_pad's grabs. And plug in
the default grab.
Features like "holding" the focus should be implemented via grabs, like
they are for pointer and keyboard.
The override_redirect flag can change on configure notify and
on map notify. This adds an event to know when it changes.
This removes wlr_xwayland_surface_is_unmanaged which was wrongly
using the window type to decide whether the view should be
unmanaged.
A similar patch was proposed to Weston, but has never been
merged upstream [1].
[1]: https://patchwork.freedesktop.org/patch/211161/
There were a few issues after rebase, that the merge algorithm didn't
throw at my face:
wlr_output did a check on the actual role, not a string anymore, so that
had to go to allow tablet-v2 to set cursor surfaces.
A few L_DEBUG/L_ERRORs were still around
There was a user-after-free in tablet-group free()ing, probably after
insufficient testing from a previous feedback pass
The previous naming was based on the input-device capability names from
libinput.
With code that uses the libinput_tablet_tool and mapping into tablet-v2,
this is confusing, so the name is changed to follow the names used in
the protocol.
This adds the management code to manage tablet tools lifetimes from
libinput.
It follows the suggestion made in the tablet-unstable-v2.xml to destroy
tablet_tools once all tablets that it got into contact with were removed
from the system. This is implemented via a refcount.
If a tool is *not* unique, it will be destroyed on proximity out. This
is libinput specific and mentioned in libinput docs that tools will not
be found again, so we shouldn't keep a reference to them.
Also they can't be on other tablets as well, because they cannot be
tracked.
The naming in this commit is a bit off (to not break things).
The wlr names stay the same, tablet_tool is the libinput_device with
capaiblity LIBINPUT_DEVICE_CAP_TABLET_TOOL which is more akin to
"tablet" in the tablet-unstable-v2 protocol.
The struct that corresponds to the tablet_tool in tablet-unstable-v2 is
called tablet_tool_tool, a rename should be done at some point in the
future.
To begin with, no-op updates are unnecessary, so this patch is an
improvement on its own.
Then, this fixes hotplugging issues with xwayland. xwayland waits
for both wl_output and xdg_output to send a "done" event. However,
it doesn't handle well desynchronized "done" updates: if xdg-output
sends "done" twice, the second one will wait for the next wl_output
"done" event. This is an issue when the first is a no-op and the
second is a real update: the second isn't applied. I've considered
patching xwayland instead, but it seems pretty complicated.
