Implement a simple register and lookup for function tables. This is
intended for plugins to expose APIs to other plugins.
It has been very hard to arrange a plugin to be able to call into
another plugin without modifying Weston core to explicitly support each
case. This patch fixes that.
The tests all pass.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Signed-off-by: Giulio Camuffo <giuliocamuffo@gmail.com>
Place it with the other weston_seat functions.
Signed-off-by: Bryce Harrington <bryce@osg.samsung.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
The name suggests that it activates surfaces, but the code says it
rather just assigns keyboard focus. Rename it for clarity, and so the
original function name could be used for something more appropriate
later. Switch order of parameters since keyboard focus is a property of
the seat. Update all callers as appropriate.
Change was asked for by pq, May 26, 2016:
"This should be called weston_seat_set_keyboard_focus(seat, surface).
Keyboard focus is a property of the seat."
Signed-off-by: Bryce Harrington <bryce@osg.samsung.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Useful for pointing out where the image comparisons fail.
Internal-screenshot-test is modified to save the visualization if the
test fails.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Screenshooting does not involve creating a wl_surface, so using struct
surface is superfluous.
Return a struct buffer instead. It could have been just a
pixman_image_t, but setting up proper destruction would be a bit more
work. Should not hurt to keep the wl_buffer around until the user is
ready to free the image.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Daniel Stone <daniels@collabora.com>
This rewrites write_surface_as_png() into write_image_as_png(), which
operates on a pixman_image_t instead of a struct surface.
This is part of the migration to use pixman_image_t everywhere without
superfluous parameters/members.
Now the image saving handles more than just ARGB32 format, presumably.
At least it does not assume everything is always ARGB32.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Daniel Stone <daniels@collabora.com>
This rewrites load_surface_from_png() to load_image_from_png(), to
return a pixman_image_t instead of a struct surface.
A loaded image has no need for wl_buffer or wl_surface or any of the
associated attributes. This is part of unifying to make everything use
pixman_image_t.
cairo_surface_flush() is added, because Cairo documentation for
cairo_image_surface_get_data() says you have to flush after drawing,
before using the data. It is unclear if loading a PNG counts as drawing,
so stay on the safe side.
load_image_from_png() now pays attention to the pixel format returned by
Cairo, which seems to come out as CAIRO_FORMAT_RGB24 in
internal-screenshot-test, not as CAIRO_FORMAT_ARGB32 as expected. I do
not know if Cairo actually guarantees the x8/a8 channel to be 0xff for
RGB24, but better to not trust it. Therefore the image is explicitly
converted to a8r8g8b8 as needed. This also adds support for loading A8
and RGB16_565 images, provided that Cairo delivers them.
The cairo surface is now wrapped directly into a pixman_image_t. If the
pixel format conversion is not needed, this eliminates a copy of the
image data. The Cairo surface will get automatically destroyed with the
Pixman image.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Daniel Stone <daniels@collabora.com>
check_surfaces_geometry() is removed as it was not used by anything, and
unlikely would be.
check_surfaces_equal() is merged into check_surfaces_match_in_clip(),
passing a NULL clip means to compare whole images.
check_surfaces_match_in_clip() is converted to work on pixman_image_t
instead of struct surface. The function is only concerned about
comparing images in memory, and does not care about a wl_buffer or a
wl_surface.
The verbosity of image comparisons is greatly reduced. An image mismatch
no longer prints a flood of raw pixel values. This will be replaced
later with a function writing out an error image instead.
Degenerate comparisons are no longer accepted, be that clip outside
images or zero area. Those are an indication of a programmer error.
The pixel format assumptions are made more visible in the code.
A new internal helper image_check_get_roi() computes and verifies the
area to be compared. Image iterator helper makes it simpler to write
manual pixel-poking loops.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Change create_shm_buffer() to handle any pixel format known to Pixman.
Presumably in the future we might want to test e.g. RGB565 content with
screenshot tests.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Daniel Stone <daniels@collabora.com>
No users remain outside the file. This will allow to fix the assumptions
in the function.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Daniel Stone <daniels@collabora.com>
This removes the uses of create_shm_buffer() from this test.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Daniel Stone <daniels@collabora.com>
This removes the uses of create_shm_buffer() from this test.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Daniel Stone <daniels@collabora.com>
This removes the uses of create_shm_buffer() from this test.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Daniel Stone <daniels@collabora.com>
We are growing more tests that need to handle buffers, both just images
and wl_buffers. Particularly the screenshooting facility needs these.
Currently everything is in struct surface, which contains more than we
need. It is a bit messy.
Create a new struct buffer to encapsulate the image representation, the
wl_buffer, and enough information to tear it all down (munmap) so we
don't have to leak everything. Some tests might start doing things in
loops, and leaking would accumulate.
Instead of inventing our own image representation, use pixman_image_t.
It is a well-tested library worth using, and we already rely on it in
other places.
This makes the tests depend on Pixman, which requires the fix for
building buffer-count, which would otherwise not find pixman.h.
The new create_shm_buffer_a8r8g8b8() creates an image with an explicit
format, and pixman_image_t keeps track of it. And stride and size and
data. This implementation is still a little hacky due to calling
create_shm_buffer().
A very new thing is buffer_destroy(). Previously we didn't really free
any buffers. It is not a problem when the process will exit soon anyway,
but it may become a problem if tests start iterating things.
Manual memset() on a image is converted to a pixman action, just to show
how to do it properly with pixman.
Stride and pixel format assumptions still linger all around, but those
are for another patch.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Daniel Stone <daniels@collabora.com>
A surface can be added to many layers.
This test is implemented to test this use-case
and the correct behaviour of get_layers_under_surface
API.
Signed-off-by: Emre Ucan <eucan@de.adit-jv.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
This patch introduces ivi_layout_view data struct,
which is a wrapper of weston_view.
There is always only one ivi_layout_view for an
ivi_layout_surface and ivi_layout_layer pair.
A surface could have many views with different
geometry and transformations, so that a surface
can be shown on:
1. On many screens
2. On the same screen with different positions
The geometry of a view is modified, when properties of
its layer are changed through ivi_layout_interface.
Users of ivi_layout_interface does not have direct access
to ivi_layout_view structure.
v2 changes:
1. Use ivi_view_is_rendered function instead of
active member
2. Add descriptions to introduced members of
structs
Signed-off-by: Emre Ucan <eucan@de.adit-jv.com>
[Natsume: add empty line]
Reviewed-by: Wataru Natsume <wnatsume@jp.adit-jv.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Add surface_created listener after the initialization of launchers.
Otherwise, surfaces of the launchers will be added to the application
layer too.
This does create a race where we might miss some surfaces that get
created before the UI client signals ready, but it was agreed the race
is not significant. You cannot use the launchers before the UI is ready,
and someone using systemd integration to launch clients in parallel to
Weston with ivi-shell and hmi-controller is unlikely. After all,
hmi-controller is just a demo.
Signed-off-by: Emre Ucan <eucan@de.adit-jv.com>
[Pekka: added extra commit message notes]
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Currently, the gl-renderer setup is being done on per-output
basis. This isn't desirable when trying to make weston run
with zero outputs.
When there are no outputs present, there is no surface available
to attach an EGLContext to with eglMakeCurrent, which makes
any EGL command fail.
The problem is solved by using EGL_KHR_surfaceless_context to
bind an EGLContext to EGL_NO_SURFACE, or if that is
unavailable, creating a dummy PbufferSurface and binding an
EGLContext to it, so EGL gets set up properly.
v2:
- Move PbufferSurface creation into its own function
- Introduce a new EGLConfig with EGL_PBUFFER_BIT set
and use it to create a PbufferSurface
- Make PbufferSurface attributes definition static
- Check for return of gl_renderer_setup and terminate
in case it fails
- Remove redundant gl_renderer_setup call from
gl_renderer_output_create
- Only destroy the dummy surface if it is valid
This patch causes a warning from Mesa when using the i965 driver:
libEGL warning: FIXME: egl/x11 doesn't support front buffer rendering.
A bug has been filed about it since it seems to be spurious:
https://bugs.freedesktop.org/show_bug.cgi?id=96694
Signed-off-by: Armin Krezović <krezovic.armin@gmail.com>
[Pekka: filed a Mesa bug and added the note in commit msg]
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Currently, the keyboard client is created and the input
panel surface is set as toplevel on the first output it
finds. This does not work in a scenario when there are
no outputs, resulting in weston-keyboard to crash at
startup due to operating on an invalid output pointer.
This makes input panel toplevel setting depend on a
valid output, and if there was no output present at
startup, it will be set toplevel as soon as an output
gets plugged in.
v2:
- Remove dependency on output pointer at startup
- Only setup output_configure_handler after the
keyboard has been created
- Let the output_configure_handler handle toplevel
setting in all cases
Signed-off-by: Armin Krezović <krezovic.armin@gmail.com>
Reviewed-by: Quentin Glidic <sardemff7+git@sardemff7.net>
[Pekka: fixed a line break]
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
When there are no outputs present, an output pointer
can be NULL. Dereferencing such pointer will result
in a crash.
Signed-off-by: Armin Krezović <krezovic.armin@gmail.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
When there are no outputs present, an output pointer
can be NULL. Dereferencing such pointer will result
in a crash.
Reviewed-by: Quentin Glidic <sardemff7+git@sardemff7.net>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Signed-off-by: Armin Krezović <krezovic.armin@gmail.com>
Currently, display_get_output returns a first member
of the linked list, which can never be NULL.
This is problematic, as the function would return a
dangling pointer and NULL pointer checks wouldn't
work where needed and some of the invalid members
would get accessed that way, resulting in a crash.
Reviewed-by: Quentin Glidic <sardemff7+git@sardemff7.net>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Signed-off-by: Armin Krezović <krezovic.armin@gmail.com>
Currently, get_default_output returns a first member
of the linked list, which can never be NULL.
This is problematic, as the function would return a
dangling pointer and NULL pointer checks wouldn't
work where needed and some of the invalid members
would get accessed that way, resulting in a crash.
Reviewed-by: Quentin Glidic <sardemff7+git@sardemff7.net>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Signed-off-by: Armin Krezović <krezovic.armin@gmail.com>
This patch adds a new command line option which can be
used to tell headless backend not to create any
virtual outputs.
This will be used for output hotplug emulation, where
weston will start with no outputs available, and the
virtual output will be created at runtime.
v2:
- Use bool instead of int for the indicator flag
- Move final newspace to a separate line in command
line options
Reviewed-by: Quentin Glidic <sardemff7+git@sardemff7.net>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Signed-off-by: Armin Krezović <krezovic.armin@gmail.com>
sprintf can overflow the fixed length title which is char[32]. This
patch change title to dynamically allocated char array using asprintf or
strdup. If one of them fail we leave returning NULL to indicate the
failure.
Signed-off-by: Benoit Gschwind <gschwind@gnu-log.net>
Reviewed-by: Yong Bakos <ybakos@humanoriented.com>
Tested-by: Yong Bakos <ybakos@humanoriented.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
Signed-off-by: Daniel Stone <daniels@collabora.com>
weston-tests-env is a beast to handle, when you would like to start
weston manually for a test you wan to start inside gdb. This patch
causes the full command line to be printed to the automake test logs, so
you can copy it from there and run it by hand.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Quentin Glidic <sardemff7+git@sardemff7.net>
Write the command line to the log to aid debugging. It needs to be
copied before parsing, because parsing mutates argv.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Quentin Glidic <sardemff7+git@sardemff7.net>
This fix also depends on "compositor-headless: do not create a seat".
If we lose the race against weston-desktop-shell setting cursors, which
is very rare, we get a cursor image in the screenshot, causing the test
to fail. This is now fixed by moving the (remaining) cursor out of the
way.
Arguably we should have better solutions for this, but that is another
story. This is a stop-gap measure we can copy also in new
screenshooting tests.
v2: Remove the example code for how to trigger the race, and rewrite the
big comment.
Cc: Derek Foreman <derekf@osg.samsung.com>
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>
This clarifies what is supposed to be the libweston code.
v2: screen-share.c is already in compositor/ instead.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Yong Bakos <ybakos@humanoriented.com>
Acked-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Quentin Glidic <sardemff7+git@sardemff7.net>
Tested-by: Quentin Glidic <sardemff7+git@sardemff7.net>
Tested-by: Benoit Gschwind <gschwind@gnu-log.net>
Acked-by: Benoit Gschwind <gschwind@gnu-log.net>
[Pekka: rebased]
This is the start of separating weston-the-compositor source files from
libweston source files.
This is moving all the files related to the 'weston' binary. Also the
CMS and systemd plugins are moved.
xwayland plugin is not moved, because it will be turned into a
libweston feature.
To avoid breaking the build, #includes for weston.h are fixed to use
compositor/weston.h. This serves as a reminder that such files may need
further attention: moving to the right directory, or maybe using the
proper -I flags instead.
v2: Move also screen-share.c, and add a note about weston-launch.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Yong Bakos <ybakos@humanoriented.com>
Acked-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Quentin Glidic <sardemff7+git@sardemff7.net>
Tested-by: Quentin Glidic <sardemff7+git@sardemff7.net>
Tested-by: Benoit Gschwind <gschwind@gnu-log.net>
Acked-by: Benoit Gschwind <gschwind@gnu-log.net>
[Pekka: rebased]
The headless backend was creating a seat, but nothing would ever happen
on that seat, except the pointer cursor would show up.
Tests do not seem to use this seat either. weston-test.so plugin creates
its own seat to be controlled by the tests.
This is part of the series to get rid of cursors for screenshot-based
tests where they are not wanted.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>
Check that the keyboard init in weston-test.so plugin succeeds.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>
At the moment only XRGB8888 is supported when using pixman renderer.
This patch adds support also for RGB565.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>
Signed-off-by: Daniel Stone <daniels@collabora.com>
Pixman's headers include a representation of -1 in fixed-point, which is
-1 << 16. This trips a GCC warning about shifting negative values. As we
can't do much about it, just silence the warning.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>
compositor->config was removed a while ago.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Quentin Glidic <sardemff7+git@sardemff7.net>
Use three buffers like simple-dmabuf-v4l instead of just two.
This is required, because when a frame callback arrives, the just
committed buffer is only on its way to the screen, while the previous
buffer is still being scanned out. It will take for the page flip to
complete, before the previous buffer is release. However, we want to be
able to repaint already at the frame callback, so three buffers can be
necessary.
This patch fixes weston-simple-dmabuf-intel to not abort with "Both
buffers busy at redraw()." when hardware overlays are used and the
surface gets directly scanned out.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Yong Bakos <ybakos@humanoriented.com>
For debugging weird timing issues. If your clock resolution is
unexpectedly e.g. 10 ms, you can be sure you will have strange timing
issues. This is almost certainly caused by kernel misconfiguration.
We rely on clock_getres() being available by the same thing that gets us
clock_gettime(), so that no new configure.ac check is needed.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-By: David Fort <contact@hardening-consulting.com>
EGL code was added to the fbdev backend in
4aa756dc7a in 2013, apparently for running
Weston on libhybris with Android hardware drivers.
This actually had nothing to do with the fbdev backend, really. Fbdev
was just a convenient platform to plug in the EGL init code and load
GL-renderer. Fbdev itself was not used at all in this case.
Fbdev should be forgotten and die, as we have better interfaces for
accelerated rendering and for controlling displays. It may be a bit too
harsh to remove the whole fbdev backend just yet, but let us at least
simplify it this much.
Fbdev+EGL has been the unholy union used by proprietary driver stacks of
hardware vendors in the non-PC world as a quick and dirty way to get
something out on the screen. In these cases it is actually the EGL
implementation that does everything internally, fbdev is not needed.
Fbdev was never meant for the sort anyway.
If anyone still needs this use case, I recommend sticking with a
outdated Weston to match your outdated platform. Or if you really
insist, write a new backend that does not pretend to use fbdev and just
initializes EGL and GL-renderer.
Cc: Adrian Negreanu <groleo@gmail.com>
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Quentin Glidic <sardemff7+git@sardemff7.net>
Reviewed-by: Derek Foreman <derekf@osg.samsung.com>
This was a left-over from ca52b31d3f.
Reported-by: Emmanuel Gil Peyrot <emmanuel.peyrot@collabora.com>
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
A surface could have more than one views.
Therefore, it is not possible to map a surface to
a specific view. The implementation of the API
iterates the list of views of the surface, and
returns the first found view.
It is not necessary to have this API to found
a view of the surface. Therefore, I removed the API.
Signed-off-by: Emre Ucan <eucan@de.adit-jv.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Change the output make value from "waywayland" to "wayland".
References: 90bc88c710
Signed-off-by: Yong Bakos <ybakos@humanoriented.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
When an output is resized (permanent mode switch), we should also notify the
shell client so that the panel and background fits to the new screen dimensions.
Signed-off-by: David Fort <contact@hardening-consulting.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
These tests poke the viewporter interface to ensure proper behaviour
from client perspective, without testing the rendering result.
These cases are covered:
- create viewport twice
- source rectangle invalid value errors, and unset
- destination size invalid value errors, and unset
- source causing non-integer destination size
- source inside/outside of buffer with transform, scale
- source outside NULL buffer, then getting real buffer
- source outside NULL buffer with inherited NULL
- set_source, set_destination, and destroy after the wl_surface is
destroyed
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
The revised wp_viewport spec requires that unset has to have all of x, y,
width and height -1 to be recognized.
Check for negative x and y and raise the required error. The error event
now mentions the wl_surface, too.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
wp_viewporter requires the destination size to be always in integers,
even when it is implicit from source rectangle. Emit appropriate
protocol errors.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
wp_viewporter spec requires protocol errors when the viewport is outside
the buffer area.
The viewport is checked in wl_surface.commit handler as the error needs
to be delivered as a reply to the commit, not at state apply time later
(e.g. with synchronized sub-surfaces, or at render time).
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
If a client destroyed the wl_surface before the wp_viewport, Weston core
would access freed memory, because the weston_surface pointer stored in
the wp_viewport wl_resource's user data was stale.
Fix this by setting the user data to NULL on wl_surface destruction. It
is specifically about wl_surface and not weston_surface destruction,
because this is about client-visible behaviour. Something internal might
keep weston_surface alive past the wl_surface.
Add checks to all wp_viewport request handlers.
At the same time, implement the new error conditions in wp_viewport:
calling any request except destroy must result in a protocol error.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
Pekka Paalanen