Allows the compositor to submit tokens to the pool of
currently active tokens. This can be useful when the
launcher doesn't use or support xdg-activation-v1 by
itself - e.g. when it is X11 based or use gtk_shell1.
This doesn't work if scene outputs are not used as the primary output of
scene surfaces will always be NULL.
Therefore, take a wlr_scene_output instead of separate wlr_scene and
wlr_output arguments and rename the function to
wlr_scene_output_send_frame_done().
The actual behavior of the function is unchanged.
This allows compositors to avoid sending multiple frame done events
to a surface that is rendered on multiple outputs at once. This may
also be used in the same way for presentation feedback.
wlroots picks names for all outputs, but it might be desirable for
compositor to override it.
For instance, Sway will use a headless output as a fallback in
case no outputs are connected. Sway wants to clearly label the
fallback output as such and label "real" headless outputs starting
from HEADLESS-1.
Implement a basic version of linux-dmabuf-unstable-v1 version 4.
Only default hints are implemented.
The new wlr_linux_dmabuf_feedback_v1 data structure will allow
compositors to define their own custom hints in the future. This
data structure makes it easy to describe feedback metadata.
It's converted to a "compiled" form suitable for marshalling over
the Wayland socket via feedback_compile.
This allows output commit listeners to access the newly committed
buffer. Currently wlr_output.front_buffer is used but it'll get
removed in the next commit.
DRM formats with an empty modifier list are invalid. Instead of
emptying the list, reduce it to { INVALID }.
Add a check to make sure the renderer and backend support implicit
modifiers, so that we don't fallback on e.g. Vulkan.
Closes: https://github.com/swaywm/sway/issues/6692
This allows getting a wlr_scene_output from a wlr_output. Since an
output can only be added once to a scene-graph there's no ambiguity.
This is useful for compositors using wlr_scene_attach_output_layout:
the output layout integration automatically creates a scene-graph
output for each wlr_output added to the layout.
This allows compositors to get primary formats without manually
calling wlr_output_impl.get_primary_formats.
For example, the Sway patch for linux-dmabuf feedback [1] needs
this.
[1]: https://github.com/swaywm/sway/pull/6313
Sometimes we were calling wlr_output_impl.set_cursor with a NULL
buffer, but we weren't clearing wlr_output.cursor_front_buffer.
Avoid leaving a dangling buffer behind.
Introduce a helper function output_set_hardware_cursor which calls
wlr_output_impl.set_cursor and keeps cursor_front_buffer in sync.
The implicit check to filter out LINEAR for dmabuf checked for INVALID
twice instead of checking for INVALID & LINEAR. Fix this.
Fixes: d37eb5c2ea ("linux-dmabuf-v1: filter out LINEAR if implicit")
Reported-by: Dawid Czeluśniak <czelusniakdawid@gmail.com>
If only INVALID and LINEAR are valid modifiers, we need to filter out
LINEAR since Xwayland won't be able to allocate a BO with the explicit
linear modifier on hardware that does not support explicit modifiers.
The addition of LINEAR is an internal implementation detail which
simplifies the wlroots architecture for now.
Evntually Xwayland should be fixed to filter out modifiers that are not
supported by the GBM implementation, see [1]. This could be done by
querying EGL for the supported modifiers.
[1]: https://gitlab.freedesktop.org/xorg/xserver/-/issues/1166
This allows compositors to easily add an xdg_surface to the
scene-graph while retaining the ability to unconstraint popups
and decide their final position.
Compositors can handle new popups with the wlr_xdg_shell.new_surface
event, get the parent scene-graph node via wlr_xdg_popup.parent.data,
create a new scene-graph node via wlr_scene_xdg_surface_tree_create,
and unconstraint the popup if they want to.
The parameters are used when the client is in the process of
building a buffer. There's no reason why this internal
implementation detail should be exposed in our public header.
All graphics drivers supporting cursor planes support ARGB8888,
the default cursor format, so this fallback is almost certainly
unused.
Essentially all cursor themes use alpha transparency to make it
clearer where relative to the screen content the cursor hotspot is.
It is better to fall back to a slightly slower software cursor than
it is to fall back to the opaque square that is a hardware cursor
without an alpha channel.
This change introduces new double buffered state to the wlr_output,
corresponding to the buffer format to render to.
The format being rendered to does not control the bit depth of colors
being sent to the display; it does generally determine the format with
which screenshot data is provided. The DRM backend _may_ sent higher
bit depths if the render format depth is increased, but hardware and
other limitations may apply.
Most (and possibly all) compositors using wlroots only ever render
fully opaque content. To provide better performance, this change
switches the default format used by wlr_output buffers from
ARGB8888 to the opaque XRGB8888.
Compositors like mutter, kwin, and weston already default to
XRGB8888, so this change is unlikely to expose any new bugs in
underlying drivers and hardware.
This does not affect the hardware cursor's buffer format, which is
still ARGB8888 by default.
As part of this change, the X11 backend (which does not support
changing format at runtime) now picks a true color, 24 bit depth
visual (i.e. XRGB8888) instead of a 32 bit depth (ARGB8888) one.
This makes it possible for the two functions using output_pick_format
(output_pick_cursor_format and output_create_swapchain) to select
different buffer formats.
The backend and renderer don't directly interact together, so there's
no point in checking that their buffer caps intersect. What we want to
check is that:
- The backend and allocator buffer caps are compatible, because the
backend consumes buffers to display them.
- The renderer and allocator buffer caps are compatible, because the
renderer imports buffers to sample them or render to them.
For instance, when running with the DRM backend and the Pixman renderer,
the (backend & renderer) check will fail because backend = DMABUF and
renderer = DATA_PTR.
They are never used in practice, which makes all of our flag
handling effectively dead code. Also, APIs such as KMS don't
provide a good way to deal with the flags. Let's just fail the
DMA-BUF import when clients provide flags.
We were send a protocol error if INTERLACED or BOTTOM_FIRST was
set. This is incorrect for the zwp_linux_dmabuf_params.create
code-path because this kills the client without allowing it to
gracefully handle the error.
We should only send a protocol error if the client provides a bit
not listed in the protocol definition.
The protocol uses a signed integer here, which is also what the
wlr_input_method_v2_preedit_string struct provides to compositors from
the input method protocol. Sway currently just passes those int32_t
values directly to this function leading to an implicit conversion.
The data field is useful to track metadata about a token. The destroy
events are useful for compositors that track application startup to
let them know they can stop doing that.
These new functions allow a compositor to request new managed tokens
without participating in the xdg-activation procedure as a wayland
client.
This enables the compositor itself to behave as a launcher
application.
Variables on the stack are released when the parent block is closed.
Here, `now` is used outside of the `if` block, causing the following
crash when starting Sway with the headless backend:
==49606==ERROR: AddressSanitizer: stack-use-after-scope on address 0x7fff94645f90 at pc 0x5558aeae9e29 bp 0x7fff94645df0 sp 0x7fff94645de0
READ of size 16 at 0x7fff94645f90 thread T0
#0 0x5558aeae9e28 in handle_present ../sway/desktop/output.c:834
#1 0x7fdc8d6792fb in wlr_signal_emit_safe ../subprojects/wlroots/util/signal.c:29
#2 0x7fdc8d54f77f in wlr_output_send_present ../subprojects/wlroots/types/output/output.c:766
#3 0x7fdc8d524a28 in output_commit ../subprojects/wlroots/backend/headless/output.c:71
#4 0x7fdc8d54d2db in wlr_output_commit ../subprojects/wlroots/types/output/output.c:629
#5 0x5558aeb013cb in output_render ../sway/desktop/render.c:1157
#6 0x5558aeae549e in output_repaint_timer_handler ../sway/desktop/output.c:544
#7 0x5558aeae5f8a in damage_handle_frame ../sway/desktop/output.c:606
#8 0x7fdc8d6792fb in wlr_signal_emit_safe ../subprojects/wlroots/util/signal.c:29
#9 0x7fdc8d6007d5 in output_handle_frame ../subprojects/wlroots/types/wlr_output_damage.c:44
#10 0x7fdc8d6792fb in wlr_signal_emit_safe ../subprojects/wlroots/util/signal.c:29
#11 0x7fdc8d54ee84 in wlr_output_send_frame ../subprojects/wlroots/types/output/output.c:720
#12 0x7fdc8d54efc3 in schedule_frame_handle_idle_timer ../subprojects/wlroots/types/output/output.c:728
#13 0x7fdc8c9dcf5a in wl_event_loop_dispatch_idle (/usr/lib/libwayland-server.so.0+0xaf5a)
#14 0x7fdc8c9dcfb4 in wl_event_loop_dispatch (/usr/lib/libwayland-server.so.0+0xafb4)
#15 0x7fdc8c9dabc6 in wl_display_run (/usr/lib/libwayland-server.so.0+0x8bc6)
#16 0x5558aeac8e30 in server_run ../sway/server.c:285
#17 0x5558aeac3c7d in main ../sway/main.c:396
#18 0x7fdc8be35b24 in __libc_start_main (/usr/lib/libc.so.6+0x27b24)
#19 0x5558aea8686d in _start (/home/simon/src/sway/build/sway/sway+0x33f86d)
If the output is destroyed after capture_output but before
frame_handle_copy, it'll have a dangling output pointer. Add the
output destroy listener in capture_output.
Closes: https://github.com/swaywm/wlroots/issues/3284
This is only called from one function.
To destroy the wlr_scene_subsurface_tree from elsewhere, callers
can destroy the scene-graph node returned by
wlr_scene_subsurface_tree_create instead (just like a compositor
would do). subsurface_tree_handle_surface_destroy does exactly this.
Inlining avoids calling subsurface_tree_destroy by mistake.
This organizes the wlr_output implementation into separate files.
This avoids having a single mega-file with lots of unrelated parts
and makes it more obvious what the interactions between all the
parts are.
No functional changes, just moving code around.
Currently these functions remove the node from the scene if the sibling
argument is the same node as the node. To prevent confusion when
misusing this API, assert that the nodes are distinct and document this.
These functions are used mostly for rendering, where including unmapped
surfaces is undesired.
This is a breaking change. However, few to no usages will have to be
updated.
struct wlr_xdg_surface_state is introduced to hold the geometry
and configure serial to be applied on next wl_surface.commit.
This commit fixes our handling for ack_configure: instead of making
the request mutate our current state, it mutates the pending state
only.
Co-authored-by: Simon Ser <contact@emersion.fr>
As touchpad touches are generally fully abstracted, a client cannot
currently know when a user is interacting with the touchpad without
moving. This is solved by hold gestures.
Hold gestures are notifications about one or more fingers being held
down on the touchpad without significant movement.
Hold gestures are primarily designed for two interactions:
- Hold to interact: where a hold gesture is active for some time a
menu could pop up, some object could be selected, etc.
- Hold to cancel: where e.g. kinetic scrolling is currently active,
the start of a hold gesture can be used to stop the scroll.
Unlike swipe and pinch, hold gestures, by definition, do not have
movement, so there is no need for an "update" stage in the gesture.
Create two structs, wlr_event_pointer_hold_begin and
wlr_event_pointer_hold_end, to represent hold gesture events and the
signals to emit them: wlr_pointer->pointer.hold_begin/hold_end.
This allows the compiler to error out if we haven't enumerated all
of the cases. This is useful to avoid a missing implementation when
adding a new node type.
This commit removes any checks whether a configure will change anything
and makes configures be sent unconditionally. Additionally, configures
are scheduled on xdg_toplevel.{un,}set_{maximized,fullscreen} events.
Previously, `wlr_xdg_toplevel` didn't follow the usual "current state +
pending state" pattern and instead had confusingly named
`client_pending` and `server_pending`. This commit removes them, and
instead introduces `wlr_xdg_toplevel.scheduled` to store the properties
that are yet to be sent to a client, and `wlr_xdg_toplevel.requested`
to store the properties that a client has requested. They have different
types to emphasize that they aren't actual states.
This allows callers to specify the operations they'll perform on
the returned data pointer. The motivations for this are:
- The upcoming Linux MAP_NOSIGBUS flag may only be usable on
read-only mappings.
- gbm_bo_map with GBM_BO_TRANSFER_READ hurts performance.
This will allow more scene-graph extensions to be added without
cluttering wlr_scene.c, for instance for sub-surface handling and
wlr_output_layout integration.
When providing non-zero layout-local coordinates to
wlr_scene_render_output, the viewport should be translated by the
given values. However the viewport was translated by the opposite
values: when giving 42,42 the viewport's position would be set to
-42,-42.
On modeset wlr_output will internally allocate a buffer. The
backend will emit a "mode" output event, then wlr_output will
emit a "commit" event.
wlr_output_damage handles the "mode" event by damaging the whole
output, and then handles the "commit" event. However the commit
event has a buffer, so wlr_output_damage rotates the damage in its
ring buffer, thinking the compositor has rendered a frame. The
compositor hasn't rendered a frame, what wlr_output_damage sees is
the internal wlr_output black buffer used for the modeset.
Let's fix this by damaging the whole output in the "commit" event
handler if the mode has changed. Additionally, damage the whole
output after rotating the damage ring buffer.
Caching frame callback lists is actually the correct behavior, because
if a surface is locked because of e.g. subsurface synchronization,
clients would expect to receive frame done events only after the
pending state is actually committed.
With the addition of a non-surface node type, it was unclear how such
nodes should interact with scene_node_surface_at(). For example, if the
topmost node at the given point is a RECT, should the function treat
that node as transparent and continue searching, or as opaque and return
(probably) NULL?
Instead, replace the function with one returning a scene_node, which
will allow for more consistent behavior across different node types.
Compositors can downcast scene_surface nodes via the now-public
wlr_scene_surface_from_node() if they need access to the surface itself.
RECT is a solid-colored rectangle, useful for simple borders or other
decoration. This can be rendered directly using the wlr_renderer,
without needing to create a surface.
For consistency with the rest of the scene-graph API, prevent detaching
a subtree by giving NULL for the new parent, and don't allow ROOT nodes
to be grafted into another tree.
If nodes are arranged in a tree rather than at a single level, then it
makes sense that there should be a way to move them to a completely
different parent in addition to moving up or down among siblings.
This allows compositors to easily enable or disable a scene-graph node.
This can be used to show/hide a surface when the xdg_surface is
mapped/unmapped.
A new wlr_scene API has been added, following the design ideas from [1].
The new API contains the minimal set of features required to make the
API useful. The goal is to design a solid fundation and add more
features in the future.
[1]: https://github.com/swaywm/wlroots/issues/1826#issuecomment-564601757
If a subsurface is being placed below a subsurface right above it, this
should be a noop. However, `node` pointed to the subsurface that was
moved, which resulted in `subsurface->parent_pending_link` being
inserted into itself, breaking parent's pending subsurface list.
This commit separates finding the requested node and getting it's `prev`
field, fixing the issue.
Similar to commit 85757665e6 ("backend/drm: Check if output is enabled
before sending frame event"), check if the output is still enabled
before sending the frame event. This fixes the bug not only for the DRM
backend, but for wayland and X11 as well.
The protocol specifies that all requests (aside from destroy) are
ignored after the compositor sends the closed event. Therefore,
destroying the wlroots object and rendering the resource inert
when sending the closed event keeps things simpler for wlroots and
compositors.
This wlr_surface_state field was a special case because we don't
want to save the whole current state: for instance, the wlr_buffer
must not be saved or else wouldn't get released soon enough.
Let's just inline the state fields we need instead.
wl_fixed_t is a 32-bit data type, but our doubles are 64-bit. This meant
that two doubles that would map to the same wl_fixed_t could compare
unequal, and send a duplicate motion event.
Refs swaywm/sway#4632.
The protocol allows compositors to not send any keymap to Wayland
clients. Handle a keymap-less keyboard correctly by sending
WL_KEYBOARD_KEYMAP_FORMAT_NO_KEYMAP instead of erroring out in the
mmap call.
When enabling an output, skip the empty buffer allocation if the
backend accepts modesets without a buffer.
This fixes mode-setting with the noop backend.
According to the viewport protocol, upon wp_viewport::destroy():
> The associated wl_surface's crop and scale state is removed.
> The change is applied on the next wl_surface.commit.
Therefore, wp_viewport_destroy(viewport) should remove all viewport state.
Currently, wlroots does not remove the crop and scale state. Instead, a
client must do:
wl_fixed_t clear = wl_fixed_from_int(-1);
wp_viewport_set_source(viewport, clear, clear, clear, clear);
wp_viewport_set_destination(viewport, -1, -1);
wp_viewport_destroy(viewport);
This commit adds the necessary logic into viewport_destroy and makes
wlroots comply with the protocol.
Sometimes we allocate a buffer with modifiers but then fail to
perform a modeset with it. This can happen on Intel because of
bandwidth limitations. To mitigate this issue, it's possible to
re-allocate the buffer with modifiers.
Add the logic to do so in wlr_output.
Adds `wlr_buffer_resource_interface` and `wlr_buffer_register_resource_interface`,
which allows a user to register a way to create a wlr_buffer from a specific
wl_resource.
The first time wlr_buffer_from_resource is called with a wl_buffer
resource that originates from wl_shm, create a new
wlr_shm_client_buffer as usual. If wlr_buffer_from_resource is called
multiple times, re-use the existing wlr_shm_client_buffer.
This commit changes how the wlr_shm_client_buffer lifetime is managed:
previously it was destroyed as soon as the wlr_buffer was released.
With this commit it's destroyed when the wl_buffer resource is.
Apart from de-duplicating wlr_shm_client_buffer creations, this allows
to easily track when a wlr_shm_client_buffer is re-used. This is useful
for the renderer and the backends, e.g. the Pixman renderer can keep
using the same Pixman image if the buffer is re-used. In the future,
this will also allow to re-use resources in the Wayland and X11 backends
(remote wl_buffer objects for Wayland, pixmaps for X11).
When wlr_output manages its own swap-chain, there's no need to
hook into the backend to grab DMA-BUFs. Instead, maintain a
wlr_output.front_buffer field with the latest committed buffer.
This function doesn't need the wl_resource anymore.
In the failure paths, wlr_buffer_unlock in surface_apply_damage
will take care of sending wl_buffer.release.
We often juggle between wlr_buffer and wlr_client_buffer variables.
Use a consistent naming: "buffer" for wlr_buffer and "client_buffer"
for wlr_client_buffer.
`wlr_client_buffer_import` is splitted in two distincts function:
- wlr_buffer_from_resource, which transforms a wl_resource into
a wlr_buffer
- wlr_client_buffer_create, which creates a wlr_client_buffer
from a wlr_buffer by creating a texture from it and copying its
wl_resource
[1] and [2] have introduced new wl_array usage in wlroots, but
contains a mistake: wl_array_for_each iterates over pointers to
the wl_array entries, not over entries themselves.
Fix all wl_array_for_each call sites. Name the variables "ptr"
to avoid confusion.
Found via ASan:
==148752==ERROR: AddressSanitizer: attempting free on address which was not malloc()-ed: 0x602000214111 in thread T0
#0 0x7f6ff2235f19 in __interceptor_free /build/gcc/src/gcc/libsanitizer/asan/asan_malloc_linux.cpp:127
#1 0x7f6ff1c04004 in wlr_tablet_destroy ../subprojects/wlroots/types/wlr_tablet_tool.c:24
#2 0x7f6ff1b8463c in wlr_input_device_destroy ../subprojects/wlroots/types/wlr_input_device.c:51
#3 0x7f6ff1ab9941 in backend_destroy ../subprojects/wlroots/backend/wayland/backend.c:306
#4 0x7f6ff1a68323 in wlr_backend_destroy ../subprojects/wlroots/backend/backend.c:57
#5 0x7f6ff1ab36b4 in multi_backend_destroy ../subprojects/wlroots/backend/multi/backend.c:57
#6 0x7f6ff1ab417c in handle_display_destroy ../subprojects/wlroots/backend/multi/backend.c:124
#7 0x7f6ff106184e in wl_display_destroy (/usr/lib/libwayland-server.so.0+0x884e)
#8 0x55cd1a77c9e5 in server_fini ../sway/server.c:218
#9 0x55cd1a77893f in main ../sway/main.c:400
#10 0x7f6ff04bdb24 in __libc_start_main (/usr/lib/libc.so.6+0x27b24)
#11 0x55cd1a73a7ad in _start (/home/simon/src/sway/build/sway/sway+0x33a7ad)
0x602000214111 is located 1 bytes inside of 16-byte region [0x602000214110,0x602000214120)
freed by thread T0 here:
#0 0x7f6ff2235f19 in __interceptor_free /build/gcc/src/gcc/libsanitizer/asan/asan_malloc_linux.cpp:127
#1 0x7f6ff1c04004 in wlr_tablet_destroy ../subprojects/wlroots/types/wlr_tablet_tool.c:24
#2 0x7f6ff1b8463c in wlr_input_device_destroy ../subprojects/wlroots/types/wlr_input_device.c:51
#3 0x7f6ff1ab9941 in backend_destroy ../subprojects/wlroots/backend/wayland/backend.c:306
#4 0x7f6ff1a68323 in wlr_backend_destroy ../subprojects/wlroots/backend/backend.c:57
#5 0x7f6ff1ab36b4 in multi_backend_destroy ../subprojects/wlroots/backend/multi/backend.c:57
#6 0x7f6ff1ab417c in handle_display_destroy ../subprojects/wlroots/backend/multi/backend.c:124
#7 0x7f6ff106184e in wl_display_destroy (/usr/lib/libwayland-server.so.0+0x884e)
previously allocated by thread T0 here:
#0 0x7f6ff2236279 in __interceptor_malloc /build/gcc/src/gcc/libsanitizer/asan/asan_malloc_linux.cpp:145
#1 0x7f6ff1066d03 in wl_array_add (/usr/lib/libwayland-server.so.0+0xdd03)
[1]: https://github.com/swaywm/wlroots/pull/3002
[2]: https://github.com/swaywm/wlroots/pull/3004
The wl_touch.frame event is used to group multiple touch events
together. Instead of sending it immediately after each touch event,
rely on the backend to send it (and on the compositor to relay it).
This is a breaking change because compositors now need to manually
send touch frame events instead of relying on wlr_seat to do it.
When wlr_output.swapchain is used instead of the backend's, the
buffer_type will be set to SCANOUT even if wlr_output_attach_render
has been called. This tricks wlr_output_damage into thinking the
whole output needs to be repainted.
Workaround this issue by forcing buffer_type to RENDER when the
output has a back-buffer set.
Will clean all of that up when removing the precommit event handler
altogether.
This commit fixes damage tracking on the Wayland, X11 and headless
backends.
Right now we rely entirely on implicit sync for synchronizing
access to GPU buffers. Implicit sync works by setting
synchronization points on the buffer in writers, and letting
readers wait on these sync points before accessing the buffer.
With OpenGL, sync points are created using functions such as
eglSwapBuffers or glFlush. If none of these special functions
are called, no sync point will be created and readers will
potentially access a buffer that hasn't finished rendering yet.
In the context of wlroots, OpenGL is the writer and the backend
(KMS or parent Wayland/X11 session) is the reader. After we're
done rendering a frame, and before passing that frame to the
backend, we need to call glFlush.
glFlush is called when the buffer is detached from the renderer.
This is a task done by output_clear_back_buffer. So let's call
this function before invoking the impl->commit hook, instead of
calling it after.
All of this is maybe a little tricky to get right with the
current renderer_bind_buffer API. The new
wlr_renderer_begin_with_buffer API is much better, because glFlush
is called on wlr_renderer_end, so it's more intuitive.
Closes: https://github.com/swaywm/wlroots/issues/3020
Everything needs to go through the unified wlr_buffer interface
now.
If necessary, there are two ways support for
EGL_WL_bind_wayland_display could be restored by compositors:
- Either by using GBM to convert back EGL Wayland buffers to
DMA-BUFs, then wrap the DMA-BUF into a wlr_buffer.
- Or by wrapping the EGL Wayland buffer into a special wlr_buffer
that doesn't implement any wlr_buffer_impl hook, and special-case
that buffer type in the renderer.
This allows renderers to choose between implementing the old
wlr_renderer_impl.texture_from_wl_drm hook, or opt for the new
wlr_drm stub. The stub has the advantage of not requiring any
special support code: stubbed wl_drm buffers look exactly like
DMA-BUFs from linux-dmabuf-unstable-v1.
This will allow us to remove all of our EGL wl_drm support code
and remove some weird stuff we need just for wl_drm support. In
particular, wl_drm buffers coming from the EGL implementation
can't easily be wrapped into a wlr_buffer properly.
We were bumping the pending sequence number after emitting the
commit event, so commit handlers were seeing inconsistent state
where current.seq == pending.seq. This prevents commit handlers
from immediately locking the pending state.
Fix this by bumping the pending sequence number before firing the
commit event.
There are still many situations where the buffer scale is not
divisible by scale. The fix will require a tad more work, so
let's just log the client error for now and continue handling
the surface commit as usual.
Closes: https://github.com/swaywm/sway/issues/6352
On commit failure, we need to unbind the back buffer from the
renderer.
This fixes assertions triggered on commits following a failed commit
where the compositor called wlr_output_attach_render.
Introduce wlr_shm_client_buffer, which provides a wlr_buffer wrapper
around wl_shm_buffer.
Because the client can destroy the wl_buffer while we still are using
it, we need to do some libwayland tricks to still be able to continue
accessing its underlying storage. We need to reference the wl_shm_pool
and save the data pointer.
This new API allows buffer implementations to know when a user is
actively accessing the buffer's underlying storage. This is
important for the upcoming client-backed wlr_buffer implementation.
Prior to this commit, subsurfaces could only be placed above their
parent. Any place_{above,below} request involving the parent would
fail with a protocol error.
However the Wayland protocol allows using the parent surface in the
place_{above,below} requests, and allows subsurfaces to be placed
below their parent.
Weston's implementation adds a dummy wl_list node in the subsurface
list. However this is potentially dangerous: iterating the list
requires making sure the dummy wl_list node is checked for, otherwise
memory corruption will happen.
Instead, split the list in two: one for subsurfaces above the parent,
the other for subsurfaces below.
Tested with wleird's subsurfaces demo client.
Closes: https://github.com/swaywm/wlroots/issues/1865
There isn't always a good time to prune old tokens. Compositors
which only implement a "give focus on activation" logic can prune
tokens on focus change. However other compositors might want to
implement other semantics, e.g. "mark urgent on activation". In this
case a focus change shouldn't invalidate other tokens.
Additionally, some tokens aren't necessarily tied to a seat.
To avoid ending up with an ever-growing list of tokens, add a timeout.
Instead of passing a wlr_texture to the backend, directly pass a
wlr_buffer. Use get_cursor_size and get_cursor_formats to create
a wlr_buffer that can be used as a cursor.
We don't want to pass a wlr_texture because we want to remove as
many rendering bits from the backend as possible.
According to libinput, release events are generated when device is unplugged,
and libinput copies this behavior for device removal. Let's do the same for
our virtual keyboard.
8f846a41fa
This is another attempt to fix#2034 and the following sway issue:
https://github.com/swaywm/sway/issues/6254
Note that we have other key repeating issues in sway, which aren't addressed
by this patch. Since the virtual keyboard itself isn't destroyed when the
keyboard grab is destroyed, we'll probably need some trick to reset the state
of the corresponding virtual keyboard when the grab is released.
https://github.com/swaywm/sway/issues/6095https://github.com/swaywm/sway/issues/6193
Previously, the same struct was used for linux-dmabuf-v1 params
and buffer. This made the whole logic a little bit awkward, because
a wlr_dmabuf_v1_buffer could either be still being constructed, or
be a complete buffer.
Introduce a separate wlr_linux_buffer_params_v1 struct for buffer
params still being constructed. Once the params are complete (ie.
once the create request is sent), the params struct is destroyed
and the buffer struct is created.
This will help with [1] as well.
[1]: https://github.com/swaywm/wlroots/issues/2664
Drop wlr_dmabuf_v1_buffer_from_params_resource and
wlr_linux_dmabuf_v1_from_resource. Contrary to wl_buffer, these
resources are internal linux-dmabuf-v1 implementation details and
should not be shared with other interfaces.
Use 128-bit hexadecimal string tokens generated with /dev/urandom
instead of UUIDs for xdg-foreign handles, removing the libuuid
dependency. Update readme and CI. Closes#2830.
build: remove xdg-foreign feature
With no external dependencies required, there's no reason not to always
build it. Remove WLR_HAS_XDG_FOREIGN as well.
Some clients (like Xwayland) will fallback to wl_drm if the compositor
doesn't explicitly advertise support for implicit modifiers, even when
the compositor supports explicit modifiers. This behavior sounds correct
from a protocol point of view.
The original motivation is the following crash:
#0 0x00007f0ddeddeef5 in raise () at /usr/lib/libc.so.6
#1 0x00007f0ddedc8862 in abort () at /usr/lib/libc.so.6
#2 0x00007f0ddedc8747 in _nl_load_domain.cold () at /usr/lib/libc.so.6
#3 0x00007f0ddedd7646 in () at /usr/lib/libc.so.6
#4 0x00007f0de033f8de in gles2_get_renderer_in_context (wlr_renderer=0x612000003640) at ../subprojects/wlroots/render/gles2/renderer.c:38
#5 0x00007f0de0341675 in gles2_begin (wlr_renderer=0x612000003640, width=3840, height=2160) at ../subprojects/wlroots/render/gles2/renderer.c:186
#6 0x00007f0de033b100 in wlr_renderer_begin (r=0x612000003640, width=3840, height=2160) at ../subprojects/wlroots/render/wlr_renderer.c:56
#7 0x00007f0de03466ed in gles2_blit_dmabuf (wlr_renderer=0x612000003640, dst_attr=0x60b000150148, src_attr=0x7fff9f9bfde0) at ../subprojects/wlroots/render/gles2/renderer.c:591
#8 0x00007f0de033d717 in wlr_renderer_blit_dmabuf (r=0x612000003640, dst=0x60b000150148, src=0x7fff9f9bfde0) at ../subprojects/wlroots/render/wlr_renderer.c:210
#9 0x00007f0de04cbc3b in frame_handle_output_commit (listener=0x611000220bc8, data=0x7fff9f9c0050) at ../subprojects/wlroots/types/wlr_screencopy_v1.c:303
The GLES2 renderer assumes it'll have a wlr_buffer to render to.
Instead of accomodating for the edge-case of rendering without a
wlr_buffer, drop wlr_renderer_blit_dmabuf calls and instead use
wlr_renderer_bind_buffer just like the rest of the wlr_renderer users.
wlr_renderer_blit_dmabuf is now unused in wlroots.
The upside is that the new blit_dmabuf function in screencopy is
renderer-agnostic.
Compute only the transform matrix in the output. The projection matrix
will be calculated inside the gles2 renderer when we start rendering.
The goal is to help the pixman rendering process.
Store the previously focused surface in `state->focused` as well as in
`state->original` when starting an implicit grab. That way at the end
of an implicit grab, the detection whether the grab started and ended
on the same surface works as intended, even if the original surface was
never left at all.
Since output_destroy() calls wl_list_remove() on the output's link,
the manager must still be valid. This is the same bug fixed in bf926e3
but with a different interface.
Value is now an enum with a new value ("on-demand") that compositors can use to allow "normal" keyboard focus semantics regardless of the layer the client surface is on. An error is sent for invalid keyboard interactivity values. The old behavior is retained for clients using the previous version of the protocol.
Also adjusted the layer-shell example program to use the new keyboard interactivity options.
This function is inferior to wlr_xdg_surface_for_each_popup_surface()
for rendering as it does not iterate over subsurfaces. Furthermore,
no compositor is known to use this to iterate popups for any purpose
other than rendering. Therefore remove the function, which may of course
be reintroduced at a later date if a use-case is found.
Instead of relying on output.pending.committed, use
wlr_output_event_commit to find out whether a buffer was committed.
Eventually output.pending will be cleared before the commit event is
emitted.
References: https://github.com/swaywm/wlroots/issues/2098
This function will allow compositors to implement input handling in a
way consistent with rendering more easily.
Calling wlr_layer_surface_v1_surface_at() and checking if the result is
a wlr_xdg_popup is flawed as there may be subsurfaces in the popup tree.
This function will allow compositors to implement input handling in a
way consistent with rendering more easily.
Calling wlr_xdg_surface_surface_at() and checking if the result is a
wlr_xdg_popup is flawed as there may be subsurfaces in the popup tree.
Since device_destroy() calls wl_list_remove() on the device's link,
the manager must still be valid. However if the manager is destroyed
before the seat as the wl_display destroy handlers are called, devices
listening for the seat to be destroyed will access free'd memory in
wl_list_remove().
Since device_destroy() calls wl_list_remove() on the device's link,
the manager must still be valid. However if the manager is destroyed
before the seat as the wl_display destroy handlers are called, devices
listening for the seat to be destroyed will access free'd memory in
wl_list_remove().
- The DRM backend initially doesn't have a frame scheduled initially.
However the compositor is expected to set a mode to start the
rendering loop (frame_pending is set to true in drm_crtc_pageflip).
- The headless and X11 backends have a timer to schedule frames, so they
ignore this hint completely.
- The Wayland backend renders a fake frame to start the rendering loop.
It's the only case where a frame is pending on init, move the
assumption there.
The DRM backend adds custom modes to wlr_output.modes
Currently modes that are added after the first occurence of
wlr_output_configuration_head_v1 are not added to wlr_output_head.mode_resources.
Only wlr_output_damage works in transformed coordinates,
wlr_output->damage is in output-buffer-local coordinates.
This essentially reverts 1ecc1b5 and fixes 1477401.
Rename wlr_renderer_get_dmabuf_formats to
wlr_renderer_get_dmabuf_texture_formats. This makes it clear the formats
are only suitable for creating wlr_textures.
wlr_surface_send_enter now stores outputs that have been entered.
Combined with a new 'bind' event on wlr_output, this allows us to delay
enter events as necessary until the respective wl_output global has been
bound.
Closes: https://github.com/swaywm/wlroots/issues/2466
xdg_popups can be destroyed by the compositor when closed. When this happens,
wlroots makes the xdg_popup surface inert and resets the xdg_surface role to
NONE.
Currently, wlroots sends a protocol error and asserts that an xdg_surface has
a role when committed. This is racy if at the same time the client commits an
xdg_popup and the compositor closes it. This patch removes the assertion and
ignores commits on xdg_surfaces without a role set.
We were previously exporting DMA-BUFs when receiving the capture_output
request, and sending a done event on wlr_output.events.precommit. Instead,
export and send done on wlr_output.events.commit.
The data of a head is only sent when it is enabled. While the head was disabled
data might have been changed. In this case clients were not informed about this
change. A later enable change that does not also update the other data must
still lead to the propagation of this data.
Since we do not know what other data was changed while the head was disabled
just send together with an enable change all current data.
The protocol requires clients to set opposing anchors when requesting
a width or height of 0.
The goal of this patch is not to break clients that rely on this
behavior but to improve the consistency of the layer shell ecosystem
through adherence to the protocol.
This brings the layer-shell api in line with that of xdg-shell and
avoids reimplementing this function in every compositor in order to
render layer shell popups correctly.
This event contains a `committed` bitfield, which allows callers to know
which output fields changed during the commit.
This allows users to setup a single atomic commit listener, instead of
setting up one listener for each event (mode, scale, transform, and so
on).
References: https://github.com/swaywm/wlroots/issues/2098
These states are distinct in the time period between the ack_configure
and the next commit on the surface. Splitting these states avoids the
following race for example:
- client starts at 1000x1000
- wlr_xdg_toplevel_set_size 500x500
- size is different -> configure sent
- client acks the configure
- wlr_xdg_toplevel_set_size 1000x1000
- compare_xdg_toplevel_state returns true since there is no pending
configure and the currently committed size is still 1000x1000
- no new configure is sent
- client commits at the size it last acked, 500x500
Because wlr_renderer_blit_dmabuf() undoes y-inversion on the source
buffer, it is incorrect to pass the y-inversion flag of the source
buffer to the user.
wlr_drag sets up keyboard, pointer and touch grabs, which block 'enter'
events (and thus focus changes). For the compositor to be able to update
focus (e.g. to focus the drop target) from the destroy handler, the
grabs must be released before the destroy event is signalled.
The output backend API is now mostly state-less thanks to the atomic
hooks (commit and test). There is one exception though: attach_render.
This function makes the rendering context current. However sometimes the
compositor might decide not to render after attach_render (e.g. when
there's nothing new to render to the back buffer). Thus
wlr_output_rollback has been introduced to revert the pending state.
Because the output backend API is mostly state-less, the only thing
wlr_output_impl.rollback needs to do is revert the current rendering
context. Rename the function to rollback_render to make this clear. Add
a check in the common wlr_output code to only call rollback_render when
attach_buffer has been previously called.
On the long term, we'll be able to remove attach_render and
rollback_render together.
During surface finalization we may not have received a new buffer,
resetting width and height in this case is wrong since we display the
old buffer in this case.
This is necessary for some grabs, which currently have no way of knowing
when the pointer/keyboard focus has left a surface. For example, without
this, a drag-and-drop grab can erroneously drop into a window that the
cursor is no longer over.
This is the plumbing needed to properly fix swaywm/sway#5220. The
existing fix, swaywm/sway#5222, relies on every grab's `enter()` hook
allowing a `NULL` surface. This is not guaranteed by the API and, in
fact, is not the case for the xdg-shell popup grab and results in a
crash when the cursor leaves a surface and does not immediately enter
another one while a popup is open (#2161).
This fix also adds an assertion to wlr_seat_pointer_notify_enter() that
ensures it's never called with a `NULL` surface. This will make Sway
crash much more until it fixes its usage of the API, so we should land
this at the same time as a fix in Sway (which I haven't posted yet).
This fixes a crash that happens when a client requests a frame on a
backend that does not implement export_dmabuf.
An assertion fails with the message:
sway: types/wlr_output.c:777: wlr_output_lock_attach_render: Assertion
`output->attach_render_locks > 0' failed.
This introduces the enter and leave events for wlr_keyboard_group.
The enter event is emitted when a keyboard is added to the group while a
key is pressed that is not pressed by any other keyboard in the group.
The data is a wl_array of the pressed key codes unique to the keyboard
that should now be considered pressed.
Similarly the leave event is emitted when a keyboard is removed from the
group while at least one key is pressed that is not pressed by any other
keyboard in the group. The data is a wl_array of the pressed key codes
unique to the keyboard that should now be considered released.
The purpose of these events are to allow the compositor to update its
state to avoid corruption. Additionally, for the leave event, the
focused surface may have been notified of a key press for some or all of
the key codes and needs to be notified of a key release to avoid state
corruption.
These were previously emitted as normal key events, but they are not
normal key events. There is no actual key press or release associated
with the events. It's purely for state keeping purposes. Emitting them
as separate events allows the compositor to handle them differently.
Since these are purely for state keeping purposes and are not associated
with an actual key being pressed or released, bindings should not be
triggered as a result of these events.
We should throw a protocol error if the relevant capability has never
existed when get_(pointer|keyboard|touch) is called. Otherwise, it
should succeed, even if the capability is not currently present.
This follows the spec, and avoids possible races with the client when
capabilities are lost.
Closes: https://github.com/swaywm/wlroots/issues/2227
When calling assert(0) instead of returning a value, -Wreturn-type
warnings are triggered because assertions can be disabled. Replace these
assertions with abort().
This is currently inconsistent with the rest of the library and a bit of
a footgun for new compositors. However, this breaks the API in a very
unfortunate way for existing compositors.
This is necessary so that sway can determine when to start emulating
pointer events -- it shouldn't start doing so during an implicit grab,
even if the pen is over a surface that doesn't bind tablet input.
Refs swaywm/sway#5302.
In the case that only one axis received an event, rotating the input can
cause the change to actually happen on the other axis, as far as clients
are concerned.
This commit updates the axes flags to be consistent post-rotation.
Fixesswaywm/sway#4776.
A client requesting frames in the ready callback may miss frames that
happen while it is waiting to receive the event or sending the request.
If that happens, the client will have an outdated frame for an
indefinite period. A new frame might not be scheduled for a very long
time.
With this change, clients will receive new frames immediately upon
request.
Check for a NULL keyboard_state.keyboard value in
seat_client_create_keyboard() before trying to use it, as is done in
other functions like seat_client_send_repeat_info(). Prevents a segfault
in certain situations on keyboard removal, as seen in the sway issue.
Closes: https://github.com/swaywm/sway/issues/5205
Closes: https://github.com/swaywm/wlroots/issues/2073
GCC is complaining about a maybe-uninitialized variable when doing a
release build. Even if that can't actually happen because all enum
values are handled, add an abort call to silence the warning.
Most of the pending output state is not forwarded to the backend prior
to an output commit. For instance, wlr_output_set_mode just stashes the
mode without calling any wlr_output_impl function.
wlr_output_impl.commit is responsible for applying the pending mode.
However, there are exceptions to this rule. The first one is
wlr_output_attach_render. It won't go away before renderer v6 is
complete, because it needs to set the current EGL surface.
The second one is wlr_output_attach_buffer.
wlr_output_impl.attach_buffer is removed in [1].
When wlr_output_rollback is called, all pending state is supposed to be
cleared. This works for all the state except the two exceptions
mentionned above. To fix this, introduce wlr_output_impl.rollback.
Right now, the backend resets the current EGL surface. This prevents GL
commands from affecting the output after wlr_output_rollback.
This patch is required for FBO-based outputs to work properly. The
compositor might be using FBOs for its own purposes [2], having leftover
FBO state can have bad consequences.
[1]: https://github.com/swaywm/wlroots/pull/2097
[2]: https://github.com/swaywm/wlroots/pull/2063#issuecomment-597614312
Check that buffer can be scanned out in wlr_output_test instead of
wlr_output_attach_buffer. This allows the backend to have access to the
whole pending state when performing the check.
This brings the wlr_output API more in line with the KMS API.
This removes the need for wlr_output_attach_buffer to return a value,
and for wlr_output_impl.attach_buffer.
Consumers call wlr_buffer_lock. Once all consumers are done with the
buffer, only the producer should have a reference to the buffer. In this
case, we can release the buffer (and let the producer re-use it).
This fixes an assertion failure if a client tries to do this, e.g. by
creating multiple toplevel objects for the same surface. If the same
role data is set multiple times, this does not cause an error, which is
how cursors use this interface.
This patch disambiguates the needs_frame event by uncoupling it from
damage. A new separate damage event is emitted when the backend damages
the output (this happens e.g. VT is changed or software cursors are
used). The event specifies the damaged region.
The wlr_output.damage field is removed. wlr_output is no longer
responsible for tracking its own damage, this is wlr_output_damage's
job.
This is a breaking change, but wlr_output_damage users shouldn't need an
update.
Bugs fixed:
- Screen flashes on VT switch
- Cursor damage issues on the X11 and headless backends
Closes: https://github.com/swaywm/sway/issues/5086
This way, wlr_output_schedule_frame will always be followed by a
wlr_output_commit. This forces the compositor to render an extra
frame before stopping the rendering loop.
To test, run wleird's frame-callback [1], make sure it's the only
visible client on the output and check the interval between frame
events is the output's refresh period instead of zero.
[1]: https://github.com/emersion/wleird/blob/master/frame-callback.c
Closes: https://github.com/swaywm/wlroots/issues/2051
Split out the client/resource handling out of wlr_buffer by introducing
wlr_client_buffer. Make wlr_buffer an interface so that compositors can
create their own wlr_buffers (e.g. backed by GBM, like glider [1]).
[1]: c66847dd1c/include/gbm_allocator.h (L7)
This function allowed backends to provide a custom function for frame
scheduling. Before resuming the rendering loop, the DRM and Wayland
backends would wait for vsync.
There isn't a clear benefit of doing this. The only upside is that we
get more stable timings: the delay between two repaints doesn't change too
much and is close to a mutliple of the refresh rate.
However this introduces latency, especially when a client misses a
frame. For instance a fullscreen game missing vblank will need to wait
more than a whole frame before being able to display new content. This
worst case scenario happens as follows:
- Client is still rendering its frame and cannot submit it in time
- Deadline is reached
- Compositor decides to stop the rendering loop since nothing changed on
screen
- Client finally manages to render its frame, submits it
- Compositor calls wlr_output_schedule_frame
- DRM backend waits for next vblank
- The wlr_output frame event is fired, compositor draws new content on screen
- On the second next vblank, the new content reaches the screen
With this patch, the wlr_output frame event is fired immediately when
the client submits its late frame.
This change also makes it easier to support variable refresh rate, since
VRR is all about being able to present too-late frames earlier.
References: https://github.com/swaywm/wlroots/issues/1925
This makes sure the `wl_list_remove(&output_power->link)` in
`output_power_destroy()` does not crash even when the output_power never
got added to a list. This can e.g. happen in the `mgmt->output ==
output` error path of `output_power_manager_get_output_power`.
Leave positioner inverted on the individual axis if it's no longer
constrained. Otherwise constraint adjustment like `slide_x & flip_y`
could render popup outside of the screen when both axes are constrained.
FixesAlexays/Waybar#532
Port back style and naming improvements suggested in
https://github.com/swaywm/wlroots/pull/2026 for
keyboard-shortcuts-inhibit. These are all internal to the implementation
and therefore unproblematic.
Also, retrieve the inhibitor resource version from the manager resource
instead of setting it statically.
Signed-off-by: Michael Weiser <michael.weiser@gmx.de>
The keyboard shortcuts inhibitor protocol is useful for remote desktop
and virtualization software in order to request all keyboard events to
be passed to it and (almost) none being resonded to by the compositor.
This allows the session at the other end of the remote desktop
connection or inside the virtual machine to be interacted with as usual
(e.g. Alt+Tab to switch windows on the remote system instead of
locally).
Add the wayland protocol to the meson build files.
Copy'n'search'n'replace the very similar idle inhibit protocol
implementation. This already provides all the basic functionality:
- creating and destroying inhibitors upon request by a client,
- destruction in reaction to destruction of surfaces or displays,
- a list of inhibitors to search through for existing ones as well as
- a signal to be sent to the compositor upon registration of a new
inhibitor.
Beyond that we add the active and inactive events to be sent to the
client and wire those to activate and deactivate functions for the
compositor to call in confirmation of activation of a new inhibitor or
(un-)suspending of an existing inhibitor e.g. in response to a special
key combination entered by the user as suggested by the protocol.
As mandated by the protocol, we check the existance of an inhibitor for
a given surface and seat upon creation and return the error provided by
the protocol for that purpose.
Signed-off-by: Michael Weiser <michael.weiser@gmx.de>
Closes: https://github.com/swaywm/wlroots/issues/1817
Previously, if the current configuration contains an output X which is
destroyed, its head is automatically removed. If the compositor submits
the new configuration after X was removed, the current output
configuration is incorrectly detected to be the same as the previous
one, and no done event is sent. To prevent this, we can just keep track
of whether the current configuration is dirty, i.e whether we have sent
a done event for it.
This change ensures that wlr_output_transform_compose correctly composes
transforms when the first transform includes a rotation and the second
transform includes a flip.
The Wayland protocol specifies output transform rotations to be
counterclockwise and applied to the surface. Previously, wlroots
copied Weston and incorrectly made rotations act clockwise on
surfaces. This commit fixes that.
This change will break compositors which expect transform rotations
to be clockwise, and the rare applications that make use of surface
transforms.
Having 1.16 results in the following error when running the compositor:
2019-04-27 17:30:50 - [wayland] wl_global_create: implemented version for 'wl_seat' higher than interface version (7 > 6)
2019-04-27 17:30:50 - [sway/input/seat.c:428] seat_create:could not allocate seat
We require wayland-server >= 1.17 for wl_seat version 7.
Fixes: a671fc51d2 ("Advertise wl_seat version 7")
Fixes: a656e486f4 ("seat: fallback to v6 if libwayland 1.17 isn't available")
Most resources must not be NULL. Make it so callers need to check for
NULL explicitly. This makes it clearer in the handlers code that the
NULL wl_resource case needs to be handled, and allows callers to make a
difference between a NULL wl_resource and an inert resource.
This fixes a memory leak the refresh_state function for
wlr_keyboard_group. The event struct was being dynamically allocated and
never free'd. This changes it to a static allocation.
Since [1], the xdg-output description is mutable. Listen to output
description changes and send the new output description when updated.
[1]: 048102f21a
wlr_output.description is a string containing a human-readable string
identifying the output. Compositors can customise it via
wlr_output_set_description, for instance to make the name more
user-friendly.
References: https://github.com/swaywm/wlroots/issues/1623
Bumps minimum version to 0.51.0
- Remove all intermediate static libraries.
They serve no purpose and are just add a bunch of boilerplate for
managing dependencies and options. It's now managed as a list of
files which are compiled into libwlroots directly.
- Use install_subdir instead of installing headers individually.
I've changed my mind since I did that. Listing them out is annoying as
hell, and it's easy to forget to do it.
- Add not_found_message for all of our optional dependencies that have a
meson option. It gives some hints about what option to pass and what
the optional dependency is for.
- Move all backend subdirectories into their own meson.build. This
keeps some of the backend-specific build logic (especially rdp and
session) more neatly separated off.
- Don't overlink example clients with code they're not using.
This was done by merging the protocol dictionaries and setting some
variables containing the code and client header file.
Example clients now explicitly mention what extension protocols they
want to link to.
- Split compositor example logic from client example logic.
- Minor formatting changes
In case the texture can't be imported, release the buffer so that the
client can submit another one. In case the allocation fails, disconnect
the client.
Some globals are static and it doesn't make sense to destroy them before
the wl_display. For instance, wl_compositor should be created before the
display is started and shouldn't be destroyed.
For these globals, we can simplify the code by removing the destructor
and stop keeping track of wl_resources (these will be destroyed with the
wl_display by libwayland).
Most of the time, compositors just display the surface's current buffer
on an output. Add an helper to make it easy to support presentation-time
in this case.
The wlr_presentation_feedback struct now tracks presentation feedback
for multiple resources (but still a single surface content update). This
allows the compositor to properly send presentation events even when
there is more than one frame of latency or when it references a
surface's buffer.
Backends not supporting presentation feedback call
wlr_output_send_present with a NULL event in their commit handler. Since
the commit hasn't been applied yet, commit_seq still has its old value.
We need to increment it.
An alternative would be to move commit_seq in wlr_output_state. This
would allow to have a pending and a current commit_seq.
wlr_output_send_present could take the pending commit_seq when called
with a NULL event.
This requires functions without a prototype definition to be static.
This allows to detect dead code, export less symbols and put shared
functions in headers.
This is set to the value of wlr_output.commit_seq when the frame has
been submitted. This allows tracking presentation with more then 1 full
frame of latency.
References: https://github.com/swaywm/wlroots/issues/1917
A wlr_keyboard_group allows for multiple keyboard devices to be
combined into one logical keyboard. Each keyboard device can only be
added to one keyboard group. This helps with the situation where one
physical keyboard is exposed as multiple keyboard devices. It is up to
the compositors on how they group keyboards together, if at all.
Since a wlr_keyboard_group is one logical keyboard, the keys are a set.
This means that if a key is pressed on multiple keyboard devices, the
key event will only be emitted once, but the internal state will count
the number of devices that the key is pressed on. Likewise, the key
release will not be emitted until the key is released from all devices.
If the compositor wants access to which keys are pressed and released
on each keyboard device, the events for those devices can be listened
to, as they currently are, in addition to the group keyboard's events.
Also, all keyboard devices in the group must share the same keymap. If
the keymap's differ, the keyboard device will not be able to be added
to the group. Once in the group, if the keymap or effective layout for
one keyboard device changes, it will be synced to all keyboard devices
in the group. The repeat info and keyboard modifiers are also synced
If box->width/height is <= 0, the box doesn't contain any points, and so
there is no closest point. wlr_box_closest_point should return NAN in this
case.
In addition, we need to handle empty boxes in a few other
output-layout-related places, because outputs can have size 0x0 when
they are created or destroyed.
Since e26217c51e3a5e1d7dfc95a8a76299e056497981, touchpoints can outlive
surfaces. This works fine as long as the client stays around, but fails
horribly otherwise; therefore we have to make sure that touchpoints don't
outlive their clients.
Fixes#1788
When the surface is closed, we destroy all pending serials waiting to be
accepted. This means we need to ignore any future ack events, because we
can have the following events:
1. -> configure()
2. -> close()
3. <- ack_configure()
At point 3, wlroots will error the client because of invalid serial,
however the client hasn't processed close() yet.
This destroys the xdg popups associated with a layer surface when the
layer surface is unmapped. It does not make sense to keep the popups
open when unmapped.
From the xdg-shell specification:
If the parent is unmapped then its children are managed as
though the parent of the now-unmapped parent has become the
parent of this surface. If no parent exists for the now-unmapped
parent then the children are managed as though they have no
parent surface.
It's added to manager->input_methods list in manager_get_input_method, but
wasn't removed anywhere, leading to possible use-after-free in
wlr_input_method_manager_v2_destroy.
This adds support for xdg-output-unstable-v1 version 3, added in [1].
The xdg_output.done event is now deprecated and is replaced with
wl_output.done.
[1]: 962dd53537
This commit makes more output properties (mode, enabled, scale and transform)
atomic. This means that they are double-buffered and only applied on commit.
Compositors now need to call wlr_output_commit after setting any of those
properties.
Internally, backends still apply properties sequentially. The behaviour should
be exactly the same as before. Future commits will update some backends to take
advantage of the atomic interface. Some backends are non-atomic by design, e.g.
the X11 backend or the legacy DRM backend.
Updates: https://github.com/swaywm/wlroots/issues/1640
This prevents screencopy applications from hanging because a failed
event never got sent when the output was disconnected or disabled after
the call to buffer().
Disconnecting or disabling an output between capture_output() and
ready() could cause either a NULL dereference or an incorrect
attach_render_locks count.
The documentation for wayland-server.h says:
> Use of this header file is discouraged. Prefer including
> wayland-server-core.h instead, which does not include the server protocol
> header and as such only defines the library PI, excluding the deprecated API
> below.
Replacing wayland-server.h with wayland-server-core.h allows us to drop the
WL_HIDE_DEPRECATED declaration.
AFAIK this was always set to zero. Instead, compute wl_output mode flags on the
fly.
Technically this is a breaking change, but I don't think anybody uses this
field.
When the surface is destroyed clear it's reference but wait for the up
event to destroy the touch point via wlr_seat_touch_notify_up().
If the surface is destroyed before the up event we end up with
incomplete sequences sent to the client like
[915821.276] wl_touch@3.down(146, 2475027, wl_surface@38, 0, 236.000000, 515.000000)
[915821.608] wl_touch@3.frame()
[915821.637] wl_touch@3.motion(2475027, 0, 236.000000, 515.000000)
[915821.779] wl_touch@3.frame()
so there's never an up event. While it should be something like
[2461229.051] wl_touch@3.down(81, 3236959, wl_surface@34, 0, 218.000000, 478.000000)
[2461229.435] wl_touch@3.frame()
[2461229.484] wl_touch@3.motion(3236959, 0, 218.000000, 478.000000)
[2461229.636] wl_touch@3.frame()
[2461277.520] wl_touch@3.up(82, 3237007, 0)
[2461277.681] wl_touch@3.frame()
this confuses toolkits intepreting the next down event incorrectly. So
don't destroy the touch point too early.
This change tracks, for each wlr_seat_client, the most recent serial
numbers which were sent to the client. When the client makes a
selection request, wlroots now verifies that the serial number
associated with the selection request was actually provided to that
specific client. This ensures that the client that was most
recently interacted with always has priority for its copy selection
requests, and that no other clients can incorrectly use a larger serial
value and "steal" the role of having the copy selection.
Also, the code used to determine when a given selection is superseded
by a newer request uses < instead of <= to allow clients to make
multiple selection requests with the same serial number and have the
last one hold.
To limit memory use, a ring buffer is used to store runs of sequential
serial numbers, and all serial numbers earlier than the start of the
ring buffer are assumed to be valid. Faking very old serials is
unlikely to be disruptive.
Assuming all clients are correctly written, the only additional
constraint which this patch should impose is that serial numbers
are now bound to seats: clients may not receive a serial number
from an input event on one seat and then use that to request
copy-selection on another seat.
The backend doesn't need to handle transform changes, since everything is done
in software. In fact, all of the implementations were all identical and just
set the transform.
We could add support for hardware transforms, but:
- This would require a different field (something like hardware_transform)
- Not all combinations are possible because there often are hardware
limitations
- The Wayland protocol isn't ready for this (in particular xdg-output, see [1])
This belongs to a different patch series anyway.
[1]: https://patchwork.freedesktop.org/series/52324/
It can be surprising for callers to stash pending changes, commit, get a
failure, then set some other pending changes, commit again, and get another
failure because of the previously-pending changes.
Instead, make commit reset the pending state on failure.
In addition to `button_count`, we keep track of the current buttons
pressed just as in `wlr_keyboard`.
Add `set_add` and `set_remove` to assist with this. These functions can
only be used with values greater than 0 (such as the button/key masks
for keyboards and pointers).
Partially addresses:
- https://github.com/swaywm/wlroots/issues/1716
- https://github.com/swaywm/wlroots/issues/1593
If mmap fails, it will return MAP_FAILED not NULL. Since the error
handling was incorrectly checking for NULL, MAP_FAILED was being passed
to xkb_keymap_new_from_string, on mmap failure, causing a segfault.
This just fixes the error checking.
If an output is destroyed while an idle_done event is scheduled, it
results in the following Address Sanitizer Output:
==1469==ERROR: AddressSanitizer: heap-use-after-free on address 0x6170000bb668 at pc 0x7f49aaa0c348 bp 0x7ffed5da35b0 sp 0x7ffed5da35a0
WRITE of size 8 at 0x6170000bb668 thread T0
#0 0x7f49aaa0c347 in schedule_done_handle_idle_timer ../subprojects/wlroots/types/wlr_output.c:265
#1 0x7f49aa2f875b in wl_event_loop_dispatch_idle (/usr/lib/libwayland-server.so.0+0xa75b)
#2 0x7f49aa2f8815 in wl_event_loop_dispatch (/usr/lib/libwayland-server.so.0+0xa815)
#3 0x7f49aa2f739b in wl_display_run (/usr/lib/libwayland-server.so.0+0x939b)
#4 0x556622dadd51 in server_run ../sway/server.c:216
#5 0x556622dac25d in main ../sway/main.c:397
#6 0x7f49aa0d0ce2 in __libc_start_main (/usr/lib/libc.so.6+0x23ce2)
#7 0x556622d8d09d in _start (/usr/local/bin/sway+0x3909d)
0x6170000bb668 is located 488 bytes inside of 672-byte region [0x6170000bb480,0x6170000bb720)
freed by thread T0 here:
#0 0x7f49aabc8f89 in __interceptor_free /build/gcc/src/gcc/libsanitizer/asan/asan_malloc_linux.cc:66
#1 0x7f49aa968fc2 in drm_connector_destroy ../subprojects/wlroots/backend/drm/drm.c:829
#2 0x7f49aaa0cc52 in wlr_output_destroy ../subprojects/wlroots/types/wlr_output.c:357
#3 0x7f49aa96d2e9 in scan_drm_connectors ../subprojects/wlroots/backend/drm/drm.c:1265
#4 0x7f49aa961a59 in drm_invalidated ../subprojects/wlroots/backend/drm/backend.c:135
#5 0x7f49aaa2e1e9 in wlr_signal_emit_safe ../subprojects/wlroots/util/signal.c:29
#6 0x7f49aa98319f in udev_event ../subprojects/wlroots/backend/session/session.c:52
#7 0x7f49aa2f87f1 in wl_event_loop_dispatch (/usr/lib/libwayland-server.so.0+0xa7f1)
previously allocated by thread T0 here:
#0 0x7f49aabc95a1 in __interceptor_calloc /build/gcc/src/gcc/libsanitizer/asan/asan_malloc_linux.cc:95
#1 0x7f49aa96b7a2 in scan_drm_connectors ../subprojects/wlroots/backend/drm/drm.c:1114
#2 0x7f49aa961a59 in drm_invalidated ../subprojects/wlroots/backend/drm/backend.c:135
#3 0x7f49aaa2e1e9 in wlr_signal_emit_safe ../subprojects/wlroots/util/signal.c:29
#4 0x7f49aa98319f in udev_event ../subprojects/wlroots/backend/session/session.c:52
#5 0x7f49aa2f87f1 in wl_event_loop_dispatch (/usr/lib/libwayland-server.so.0+0xa7f1)
SUMMARY: AddressSanitizer: heap-use-after-free ../subprojects/wlroots/types/wlr_output.c:265 in schedule_done_handle_idle_timer
Shadow bytes around the buggy address:
0x0c2e8000f670: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa
0x0c2e8000f680: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa
0x0c2e8000f690: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd
0x0c2e8000f6a0: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd
0x0c2e8000f6b0: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd
=>0x0c2e8000f6c0: fd fd fd fd fd fd fd fd fd fd fd fd fd[fd]fd fd
0x0c2e8000f6d0: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd
0x0c2e8000f6e0: fd fd fd fd fa fa fa fa fa fa fa fa fa fa fa fa
0x0c2e8000f6f0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa
0x0c2e8000f700: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa
0x0c2e8000f710: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa
Shadow byte legend (one shadow byte represents 8 application bytes):
Addressable: 00
Partially addressable: 01 02 03 04 05 06 07
Heap left redzone: fa
Freed heap region: fd
Stack left redzone: f1
Stack mid redzone: f2
Stack right redzone: f3
Stack after return: f5
Stack use after scope: f8
Global redzone: f9
Global init order: f6
Poisoned by user: f7
Container overflow: fc
Array cookie: ac
Intra object redzone: bb
ASan internal: fe
Left alloca redzone: ca
Right alloca redzone: cb
Remove the idle_done idle timer when the output is destroyed
This commit introduces wlr_output_schedule_done and refactors the mechanism
used to send wl_output events to clients.
wlr_output_schedule_done schedules a wl_output.done event. This allows clients
to see wlr_output property changes as atomic.
This function is also useful for add-on interfaces like xdg_output which need
to trigger a wl_output.done event to apply their new state.
wlr_output_damage_make_current has been renamed to
wlr_output_damage_attach_render, since it's just a wrapper for
wlr_output_attach_render.
wlr_output_damage_swap_buffers has been removed completely. Instead,
wlr_output_damage now listens to successful wlr_output commits and updates its
internal state accordingly.
This updates the backend part of the output API. This is mostly renaming:
make_current becomes attach_render and swap_buffers becomes commit.
This also fixes the RDP backend to support NULL damage.
This is necessary for direct scan-out and other upcoming features. This patch
changes the output API to look like the wl_surface API.
Outputs now have some double-buffered state: the frame to be submitted
(currently only wlr_renderer frames are supported) and the damaged region.
To attach a pending frame, use wlr_output_attach_render. To set the pending
damaged region, use wlr_output_set_damage.
To submit the pending state, call wlr_output_commit. This will submit the
pending frame to the backend.
To migrate from the old API to the new one:
- Replace wlr_output_make_current calls by wlr_output_attach_render
- Replace wlr_output_swap_buffers calls by wlr_output_set_damage and
wlr_output_commit
It doesn't make sense for clients to send "accept" requests to offers that
aren't drag-and-drop. I discussed with Daniel Stone to make it a protocol
error [1] but too many clients send it (e.g. GTK+). Let's just log it for now.
[1]: https://gitlab.freedesktop.org/wayland/wayland/merge_requests/11#note_149710
* idle: enable the compositors to add custom idle timeouts
* idle: add a private constructor which also creates the resource
* idle: move resource creation to the idle implementation callback
When the proximity_in event is sent for tablet_v2 and there's already a
surface that currently has tablet (tool) focus, it should be told that
this is no longer the case, so we need to send the proximity_out event.
* Add compile args for libinput to lib_wlr_types
Fixes the build on openSUSE Tumbleweed
* Remove libinput include from wlr_tablet_v2.c
+ libinput isn't used in the file
+ Also remove libinput dependency from types/meson.build
Since the fullscreen request may be made before the toplevel's surface
is mapped, the requested fullscreen output needs to be stored so it
can be retrieved on map (along with the existing fullscreen property).
This commit makes the required changes for wlr_xdg_toplevel_v6.
Since the fullscreen request may be made before the toplevel's surface
is mapped, the requested fullscreen output needs to be stored so it
can be retrieved on map (along with the existing fullscreen property).
This commit makes the required changes for wlr_xdg_toplevel.
If there were no outputs in the output layout,
wlr_output_layout_get_box would return the box:
{
.x = INT_MIN,
.y = INT_MIN,
.width = INT_MIN - INT_MAX,
.height = INT_MIN - INT_MAX
}
which results in an integer underflow for both the width and height.
This changes the logic to have the box be all zeroes, since an empty
output layout does not have a width or height and the location of
something without a size is irrelevant so this just uses the origin.
The grab serial can be used to start a pointer grab. A button pressed event
should be used for this purpose.
Thus, we should only save the grab serial if it's the first button pressed
event we send. This commit makes it so the serial is not saved if a button is
released while another button is still pressed.
In case a tool was removed, but not yet destroyed by the client, the
tool_client's tool can be NULL. We have to check that as well in the
set_cursor handler to prevent using inert resources
This commit makes sure surface->mapped is true when the unmapped event is
emitted. This is necessary because listeners can only damage surfaces that are
mapped. This is similar to the fact that the destroy event is emitted before
any destruction is actually made.
Fixes https://github.com/swaywm/sway/issues/3568
There is no point in modesetting an output to a mode that it is already
set to. Modesetting will cause the output to briefly flicker which is
undesirable for a noop. This prevents modesetting any wlr_output,
regardless of the backend, to it's currently set mode.
data-control: separate out a data_offer struct
This is a prerequisite to adding primary selection support.
data-control: separate out data_control_source
This is a prerequisite to adding primary selection support, since that
doesn't use wlr_data_source, but rather wlr_primary_selection_source.
Update the data-control protocol
data-control: add primary selection support
Merge create_offer and create_primary_offer
Extract code into data_control_source_destroy()
Fix pointer style
Move resource neutralization to destructor
Store wl_resource in the data_offer
Extract data_offer destruction into a function
It doesn't make sense to keep popups opened when unmapped. We also need to do
so in wlr_xdg_popup_destroy so that popups are destroyed in the correct order.
This supersedes f24e17259e and
04c9ca4198. These commits were manually removing
wlr_data_source destroy handlers when starting a new drag. This is error-prone.
Instead, this commit destroys the previous source whenever we start a new drag.
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
Fixes a crash in SDL2 applications when using locked pointer.
SDL2 expects a wl_pointer v1 interface and doesn't provide a handler for
the wl_pointer::frame event. This results in a "listener function for
opcode 5 of wl_pointer is NULL" abort in wayland-client.
Tested on Xonotic v0.8.2 with libSDL2 v2.0.9. Xonotic needs to be ran
with "SDL_VIDEODRIVER=wayland xonotic-sdl" for SDL2 to use the wayland
backend.
When there aren't enough CRTCs for all outputs, we try to move a CRTC from a
disabled output to an enabled one. When this happens, the old output's state
wasn't changed, so the compositor thought it was still enabled and rendering.
This commit marks the old output as WLR_DRM_CONN_NEEDS_MODESET and sets its
current mode to NULL.
Occurs on subsequent calls to
relative_pointer_manager_v1_handle_get_relative_pointer()
Steps to reproduce:
- run rootston
- run examples/relative-pointer
- switch to relative pointer more than once
Note: if done fast enough it may take more than two switches to crash.
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.
Simon Ser
Isaac Freund
Stacy Harper
Guido Günther
David Rosca
Kirill Primak
tiosgz
Simon Zeni
Rouven Czerwinski
John Lindgren
Jonathan Wong
Manuel Stoeckl
Ronan Pigott
José Expósito
Tadeo Kondrak
Andri Yngvason
Quantum
Devin J. Pohly
Isaac Freund
Tudor Brindus
zccrs
Yuya Nishihara
Ryan Farley
Kenny Levinsen
Justus Rossmeier
Daniel Kondor
Ilia Bozhinov
Marten Ringwelski
Roman Gilg
Drew DeVault
Daniel De Graaf
Antonin Décimo
xdavidwu
Greg V
Thomas Hebb
Brian Ashworth
Kalyan Sriram
David96
Scott Anderson
Aleksei Bavshin
Michael Weiser
Dorota Czaplejewicz
Josef Gajdusek
Timidger
Alynx Zhou
György Kurucz
Ivan Molodetskikh
amingin
Markus Ongyerth
Versus Void
Sebastian Krzyszkowiak
Sebastian Krzyszkowiak
Jason Francis
Ashkan Kiani
Stuart Dilts
Ryan Walklin
Ian Fan
Geoff Greer
Connor E
John Chen
Ilya Trukhanov
Alex Maese
random human