Rebuilding regions can be an expensive operation, and we're adding more
of them. This means we need to be clever about when we actually do them.
Only dirty the paint nodes when the transform or buffer size has
actually changed, not on every commit.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Akin to the paint_node_status we already have, start also tracking a
surface dirty status. This will allow us to minimise the updates we need
to make.
Currently this is only collected, with no functional change made.
Signed-off-by: Daniel Stone <daniels@collabora.com>
If there is an opaque full-screen view with a compatible SHM client
buffer left after peeling off the client-side cursor view, bypass the
renderer and let Neat VNC read from the client buffer directly.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Instead of directly converting damage from pixman_region32_t in
global coordinates to pixman_region16_t in local coordinates,
use weston_region_global_to_output() to convert to pixman_region32_t
in local coordinates and then convert again to pixman_region16_t
in the same coordinate system, using vnc_region32_to_region16().
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
We previously had our own local variable for this, but now we can just
use the one in weston_compositor.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Until now we've only had the unadorned arithmetic functions, but they're
easy to abuse and tedious to use.
For now, we just add weston_coord_global_add/sub functions and use them
where appropriate.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
This is stored as an unadorned weston_coord internally, but with getter
functions we can put together the appropriate global or surface
coordinate.
Use them where appropriate.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
Let the GL renderer render to FBOs for RDP outputs and read the pixels
into the RDP frame buffer. This allows to run the RDP backend with the
GL renderer.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Add switch statements where renderer specific API is called to prepare
for adding GL renderer support. No functional change.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Partially revert commit 89e1831cd7 ("pixman-renderer: add
weston_renderbuffer and create/destroy interface") to bring back the
shadow_surface pixman image. The renderbuffer is only wrapped around it.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Translate damage extents used to calculate glReadPixels rectangle from
global to local coordinates.
Fixes: b1606a9f2c ("gl-renderer: support automatically downloading FBO renderbuffers")
Signed-off-by: Philipp Zabel <philipp.zabel@gmail.com>
When an output is moved, all views that are not moving with it should
cause damage where they appear in it before and after the move, and all
prior damage should move with the output.
To avoid this complexity, just damage the full output after the move.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Track damage on struct weston_renderbuffer and drop the custom damage
region from struct wayland_shm_buffer.
Pass repaint damage to wl_surface_damage() instead of accumulated
renderbuffer damage.
Signed-off-by: Philipp Zabel <philipp.zabel@gmail.com>
While the GL renderer is not able to directly render into the PipeWire buffers,
it is possible to read the rendered frame from the fbo into the PipeWire buffer.
Use the automatic download to add support for the GL renderer to the PipeWire
backend.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Instead of always initializing the Pixman renderer, make the initialization
dependent on the selected renderer. This makes it easier to add other renderers.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Make it easier to understand where ptr points to by using local variables for
the spa_buffer and spa_data.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Extract the pixman renderbuffer configuration from the add_buffer function into
a helper function to simplify the addition of the GL renderer.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Use helper function for setting up the pixman renderer to simplify the addition
of the GL renderer.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
clip_quad() is a dedicated clipping function for quads that doesn't
depend on any GL renderer internal structures. It can be moved out to
the clipper to be called by both the renderer and the clipping test
client without having to duplicate code.
Signed-off-by: Loïc Molinari <loic.molinari@gmail.com>
Since both the surface rect and the transformed quad are axis-aligned
in the simple clipping path, non-zero area detection can more
efficiently be checked post-clipping by comparing opposite edges.
Signed-off-by: Loïc Molinari <loic.molinari@gmail.com>
Add a basic check to let the clipper take the simple axis-aligned path
when nodes are solely transformed with a translation and/or a scaling.
That makes some nodes like sub-surfaces (which always have their
transform enabled) take the fast path in the common case.
Signed-off-by: Loïc Molinari <loic.molinari@gmail.com>
Move vertex clipper back to single-precision floating point
intermediates. Since positions are sent down the graphics hardware as
single-precision values, this prevents useless conversions between
single and double precision values and lets compilers fit twice as
much data into vector registers. It also removes a copy by letting the
clipper store vertices directly into the vertex buffer.
This is mostly reverting the conversion to double-precision that
happened along with the switch to the weston_coord struct for vertex
coordinates (commit 10e70bf23c).
Signed-off-by: Loïc Molinari <loic.molinari@gmail.com>
Let the graphics hardware handle the transformation from surface
position to texture coordinates. Paint nodes now have a single vertex
position attribute from which texture coordinates are derived. A new
vertex shader variant handles the transformation.
Signed-off-by: Loïc Molinari <loic.molinari@gmail.com>
The clipper transforms dirty rects to surface space before clipping.
Each dirty rect is transformed by the same matrix for each surface
rect. This change decouples the transformation and the clipping code
to transform and compute the bounding box of dirty rects just once.
Signed-off-by: Loïc Molinari <loic.molinari@gmail.com>
Clip dirty rects to surface rects in surface coordinate space.
Dispatch vertices in surface coordinates and let the graphics hardware
handle the transformations. Clipping in global coordinate space
implies a useless roundtrip on the CPU to get the clipped polygons
back in surface coordinates for the buffer transformation. Clipping in
surface coordinate space prevents that.
This might seem counter-intuitive at first because in surface space
it's the dirty rects that are clipped to axis-aligned surface rects,
while it's the opposite in global space.
The projection matrix now combines the view and the output transforms.
Signed-off-by: Loïc Molinari <loic.molinari@gmail.com>
Let the GL renderer render to FBOs and read the pixels into the Neat VNC
frame buffers. This allows to run the VNC backend with the GL renderer.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
For software backends like VNC, support downloading the FBO renderbuffer
contents via glReadPixels automatically at the end of repaint_output.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Add support for creating surfaceless outputs and rendering to FBOs.
The backend has to create FBOs with the create_fbo API and pass the
resulting weston_renderbuffer handles to repaint_output.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Since the pixman_region32_t damage is initialized unconditionally, also
finalize it unconditionally. Otherwise we leak rectangle memory when
sb->output->frame is NULL.
Reported-by: Daniel Stone <daniels@collabora.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Add a to_pixman_renderbuffer() helper to consolidate the
container_of(renderbuffer, ...) macro calls.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Remove the now unused previous/total_damage regions and the
buffer/border_damage arrays, as well as the output_get_damage and
output_rotate_damage functions.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Create dummy renderbuffers to track surface buffer damage on
demand. The renderbuffer representing the surface buffer that
is currently rendered to is inferred using buffer age.
This aligns damage tracking with the Pixman renderer and will
simplify adding FBO rendering support.
The previous/total_damage regions and the buffer/border_damage
arrays are now unused except for validation.
They can be removed next.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Split the buffer age query out into a separate function.
The following patches will replace the remainder of the
output_get_damage function.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
In some circumstances (like system load) we seem to be racing with the
threads fontconfig creates and the resource release happening inside
cleanup_after_cairo(), and we would still find a cached entry holding a
font map reference while we are on the compositor exit path --
which happens in cleanup_after_cairo() when calling
cairo_debug_reset_static_data().
This was introduced with commit 823580e070, 'backend-headless: fully
release pango and fontconfig', as a way to have a clean memory leak
report, but due to the fact we can't influence how libraries manage
their threads, the solution (for now at least) would be to just remove
it entirely.
Running it at the end in the test itself, or before calling exit(2),
while it does narrow the window, it still exhibits the cairo crash assert
related to having cached entries in that font map hash table.
Reference: https://gitlab.freedesktop.org/wayland/weston/-/issues/756
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
