| Age | Commit message (Collapse) | Author |
|---|
|
References: https://lore.kernel.org/dri-devel/20220506180216.2095060-1-jason@jlekstrand.net/
|
|
We need to wait for the pending command buffer to complete before
re-using stage buffers. Otherwise we'll overwrite the stage buffer
with new contents before the texture is fully uploaded.
|
|
We need to wait for any pending command buffer to complete before
we're able to fully destroy a struct wlr_vk_texture: the Vulkan
spec requires the VkDescriptorSet to be kept alive.
So far we've done this in vulkan_end(), after blocking until the
command buffer completes. We'll soon stop blocking, so move this
logic in get_command_buffer(), where we check which commands buffers
have completed in a non-blocking fashion.
|
|
export_imported is never used, and dmabuf_flags is already checked
in query_modifier_support().
|
|
When the DRM backend is disabled, so is the DRM lease protocol.
Prevent the header from being installed in that case.
|
|
This allows the backend to have access to the frame damage, as
reported by the scanned-out client. Some KMS drivers can make use
of it (e.g. for PSR, or optimized USB transfers in the GUD driver),
and the Wayland/X11 backends forward it to the parent compositor.
|
|
Avoids polluting wlr_output.pending with temporary state. Removes
the need to call wlr_output_rollback() in error codepaths.
|
|
vkCmdCopyBufferToImage requires that the buffer offset be a multiple
of the texel block size, which for single plane uncompressed formats
is the same as the number of bytes per pixel. This commit adds an
alignment parameter to vulkan_get_stage_span which ensures that the
provided span (and the sequence of image copy operations derived which
use it) have this alignment.
|
|
Some compositors are not interested in wlr_session, for instance
nested compositors.
Disabling wlr_session removes the udev dependency.
|
|
Instead of hand-rolling get_current_time_msec(), let's just re-use
the helper we already have in "util/time.h".
|
|
0xFFFFFFFF milliseconds is 4,294,967,295 ms so about 50 days.
A little bit too close for comfort.
Use int64_t instead of uint64_t to avoid C's implicit conversion
footguns in computations.
|
|
Instead of returning an empty multi backend, fail with a clear
error when both the DRM and libinput backends are disabled.
|
|
We have no use for these.
|
|
This allows compositors which don't use wlr_renderer to still use
wlroots' wl_shm implementation.
|
|
Since this does not have a matching _SRGB-type vulkan format, add a
new shader variant/pipeline to perform the sRGB->linear texture
conversion.
|
|
This event is useful for e.g. sending the preferred buffer scale to the
client.
|
|
This is a first step towards moving texture uploading out of
wlr_compositor.
This commit allows compositors to opt-out of the texture uploading
by passing a NULL wlr_renderer. An immediate user of this is
gamescope, which currently implements a stub wlr_renderer just to
make wlr_compositor happy.
|
|
Some compositors may want to access the original buffer in their
commit handler.
|
|
|
|
Unused.
|
|
No reason why the GLES2/Pixman renderers should depend on the
backend.
|
|
We can just forward-declare the struct instead.
|
|
We don't actually need to pull that header here.
|
|
Let's just forward-declare struct wlr_backend instead.
We need to fixup the Vulkan renderer: it needs makedev(), which
got included by chance via <wlr/backend.h> → <wlr/backend/session.h>
→ <libudev.h>.
|
|
Fixes
In file included from /usr/include/wayland-server-core.h:32,
from ../types/seat/wlr_seat_pointer.c:6:
In function 'wl_fixed_from_double',
inlined from 'wlr_seat_pointer_send_axis' at ../types/seat/wlr_seat_pointer.c:367:6:
/usr/include/wayland-util.h:641:17: error: 'low_res_value' may be used uninitialized [-Werror=maybe-uninitialized]
641 | u.d = d + (3LL << (51 - 8));
| ~~^~~~~~~~~~~~~~~~~~~
../types/seat/wlr_seat_pointer.c: In function 'wlr_seat_pointer_send_axis':
../types/seat/wlr_seat_pointer.c:329:16: note: 'low_res_value' was declared here
329 | double low_res_value;
| ^~~~~~~~~~~~~
|
|
80 cols is a bit short. Linux uses a soft 100 limit, let's do the
same.
|
|
|
|
|
|
|
|
This allows users to setup event listeners before the server is
actually started.
|
|
This can be used to know when wlr_xwayland_server decides to start
a new Xwayland process. At that point the wl_client has already
been created but the Xwayland process hasn't been started yet.
|
|
|
|
|
|
References: https://gitlab.freedesktop.org/wayland/wayland-protocols/-/merge_requests/163
|
|
Instead of hardcoding builtin resource types in
wlr_buffer_from_resource(), use the modular resource interface.
|
|
Instead of hardcoding builtin resource types in
wlr_buffer_from_resource(), use the modular resource interface.
|
|
Use wlr_compositor.h and wlr_subcompositor.h instead.
|
|
|
|
This function is not used anymore. Backends have no good reason to
damage outputs.
|
|
There's no reason why the output should be damaged here. The current
buffer doesn't need to be re-painted.
|
|
Backends no longer need to manually call wlr_output_update_enabled()
in their commit hook: wlr_output will take care of that.
|
|
References: https://gitlab.freedesktop.org/mesa/drm/-/merge_requests/272
|
|
This was previously a use-after-free in `wlr_drag.c`.
|
|
This function deals with multiple kinds of backends. Make it more
obvious that this variable holds the multi backend which is returned
to the user.
|
|
This no longer has purpose.
|
|
Up until now, wlr_backend_autocreate() created the wlr_session and
then stuffed it into struct wlr_multi_backend so that compositors
can grab it later.
This is an abuse of wlr_multi_backend and the wlr_backend API:
wlr_backend_get_session() and wlr_multi_backend.session only exist
to accomodate the needs of wlr_backend_autocreate(). What's more,
the DRM and libinput backends don't implement
wlr_backend_impl.get_session.
Instead, return the struct wlr_session to the compositor in the
wlr_backend_autocreate() call. wlr_backend_get_session() will be
removed in the next commit.
|
|
|
|
Use a basic linked list to store the currently active mappings.
Note that we don't actually need to implement a full lock-free
atomic linked list here. The signal handler will never write to
the list, it will only read it. Only the main thread will write.
We need to always expose a consistent view of the list to the
signal handler (the main thread might be interrupted at any point
by the signal handler).
|
|
It's been superseded by wlr_shm.
|
|
|
