Daniel Stone authored 6 years ago and Pekka Paalanen committed 6 years ago

This uses pip to install Meson in order to get a sufficiently new
version.

Signed-off-by: Daniel Stone <daniels@collabora.com>

Extends CI build coverage to catch more issues.

freerdp2-dev needs stretch-backports.

Signed-off-by: Pekka Paalanen <pq@iki.fi>

Extends CI build coverage to catch more issues.

Signed-off-by: Pekka Paalanen <pq@iki.fi>

This should let people using the dist tarballs to build with meson.

Signed-off-by: Pekka Paalanen <pq@iki.fi>

Daniel Stone authored 8 years ago and Pekka Paalanen committed 6 years ago

Meson is a build system, currently implemented in Python, with multiple
output backends, including Ninja and Make. The build file syntax is
clean and easy to read unlike autotools. In practise, configuring and
building with Meson and Ninja has been observed to be much faster than
with autotools. Also cross-building support is excellent.

More information at http://mesonbuild.com



Since moving to Meson requires some changes from users in any case, we
took this opportunity to revamp build options. Most of the build options
still exist, some have changed names or more, and a few have been
dropped. The option to choose the Cairo flavour is not implemented since
for the longest time the Cairo image backend has been the only
recommended one.

This Meson build should be fully functional and it installs everything
an all-enabled autotools build does. Installed pkg-config files have
some minor differences that should be insignificant. Building of some
developer documentation that was never installed with autotools is
missing.

It is expected that the autotools build system will be removed soon
after the next Weston release.

Signed-off-by: Daniel Stone <daniels@collabora.com>
Co-authored-by: Pekka Paalanen <pq@iki.fi>
Signed-off-by: Pekka Paalanen <pq@iki.fi>

Daniel Stone authored 6 years ago and Pekka Paalanen committed 6 years ago

No need to use the protocol directory prefix. This may even be necessary
for the Meson build.

Signed-off-by: Daniel Stone <daniels@collabora.com>

Extracted from the patch adding the Meson build system.

Signed-off-by: Pekka Paalanen <pq@iki.fi>

The user of backtrace() was removed in
bb707dc0 and has been unused since.

Signed-off-by: Pekka Paalanen <pq@iki.fi>

weston-egl-ext.h was moved to shared/ in
ffff92d5

Use the correct include path, so that Meson build does not need to add
../shared to the include path just for this.

Signed-off-by: Pekka Paalanen <pq@iki.fi>

Daniel Stone authored 6 years ago and Pekka Paalanen committed 6 years ago

Give it a more regular name, matching all the other test plugins.

Signed-off-by: Daniel Stone <daniels@collabora.com>

Some members have been removed from FreeRDP structs, so let's use local
variables.

This would make weston-debug much more readable when looking at
the pixel format of the buffer.

Signed-off-by: Marius Vlad <marius.vlad0@gmail.com>

In current form SHM buffers pixel format can only be printed as 0 and 1.
With the help of this helper we align with DRM_FORMAT_ pixel format.

Signed-off-by: Marius Vlad <marius.vlad0@gmail.com>

Pekka Paalanen authored 6 years ago and Marius Vlad committed 6 years ago

Use the pixel format table to parse format names. This makes the parser
recognize almost all DRM format names.

Not all formats are usable, but we rely on the use to fail
appropriately. What we can use depends on the drivers anyway.

Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>

Pekka Paalanen authored 6 years ago and Marius Vlad committed 6 years ago

Add a function to find a format description by the DRM format name. This
will be useful when parsing configuration strings.

While at it, fix the two function formattings in pixel-formats.h to
match everything else in the file.

Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>

Pekka Paalanen authored 6 years ago and Marius Vlad committed 6 years ago

There is often a need to print the name of a pixel format. Printing the
raw numeric value is hard to decipher, printing the four ASCII
characters is slightly more human-friendly but still needs a decoder
table. Add a name that can be printed easily.

The bulk of this patch was done with:
sed -i -e 's/\.format = DRM_FORMAT_\(.\+\),/DRM_FORMAT(\1),/' libweston/pixel-formats.c

Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>

It's a bit surprising that Weston looks different when launched from the root
of the git repo vs from elsewhere.

But it's also technically a security vulnerability: if I launch it from
a directory like /tmp, it might pick up a weston.ini created by another user,
which could then load modules with arbitrary code. Basically, it's the same
problem as including "." in $PATH.

Signed-off-by: Dima Ryazanov <dima@gmail.com>

eucan authored 6 years ago and eucan committed 6 years ago

input_method protocol is no longer supported.
Therefore, we should remove it from the example config

Signed-off-by: Emre Ucan <eucan@de.adit-jv.com>

eucan authored 6 years ago and eucan committed 6 years ago

input panel related members of ivi_shell struct are
not required anymore. Also get_default_view(),
input_panel_setup() and input_panel_destroy() are not used.
Therefore, we can remove them.

Signed-off-by: Emre Ucan <eucan@de.adit-jv.com>

eucan authored 6 years ago and eucan committed 6 years ago

it is unused now

Signed-off-by: Emre Ucan <eucan@de.adit-jv.com>

eucan authored 6 years ago and eucan committed 6 years ago

input panel implementation puts contents of an
application, which uses input_method protocol,
on top of all other surfaces. It is not controllable
with ivi-layout interface.

This is not acceptable for an In-Vehicle Infotainment
platform. Because we have to ensure configured scenegraph
cannot be hijacked by any rogue application.

Therefore, I am removing input panel implementation

Signed-off-by: Emre Ucan <eucan@de.adit-jv.com>

Arkadiusz Hiler authored 6 years ago and Daniel Stone committed 6 years ago

Add a couple log points for tracking atomic commits and flip processing.

Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@intel.com>

Michael Olbrich authored 6 years ago and Daniel Stone committed 6 years ago

For toplevel surfaces, the shell will do the same thing. Without this, the
commit does not trigger repaints, because the output mask for the surface
is not set.

Without this the popup is not shown unless something else triggers a
repaint. This is usually not seen because the mouse cursor triggers a
repaint at the same time.

Signed-off-by: Michael Olbrich <m.olbrich@pengutronix.de>

Vasilis Liaskovitis authored 6 years ago and Daniel Stone committed 6 years ago

Introduce a helper function to disconnect the client on unhandled
buffer types, and use it in the gl and pixman renderers. The function
is modeled after linux_dmabuf_buffer_send_server_error.

Also print the egl error state in the gl renderer, in case the
unrecognized buffer error happens when querying an egl buffer.

#148

Simon Ser authored 6 years ago and Daniel Stone committed 6 years ago

If XCURSOR_THEME or XCURSOR_SIZE is set, use it as the cursor theme
or cursor size.

This is similar to what Qt and some X11 apps do.

Deepak Rawat authored 6 years ago and Daniel Stone committed 6 years ago

Not all drivers support fb2 modifiers so read the capability before
using drmModeAddFB2WithModifiers.

Signed-off-by: Deepak Rawat <drawat@vmware.com>

If the view is not mapped, we do not need to update its
properties. We can use ivi_view_is_mapped() function to
check it.

Also we don't need to call weston_view_damage_below()
for weston_views, which were in the scenegraph. Because
we are calling weston_view_unmap for views of unmapped
ivi_views in build_view_list() function

Signed-off-by: Emre Ucan <eucan@de.adit-jv.com>

From Michael Olbrich:
"Both the core in weston_compositor_build_view_list() with view.link
and the ivi-shell in commit_screen_list() with view.layer_link
don't remove the old views from the list.

As a result, all views that are not currently in the list have
old broken links. Destroying such a view tries to remove
the view from these lists and will access the old, invalid pointers."

Therefore, we have to unmap weston_views which are not in current
scenegraph of ivi-shell. I implemented ivi_view_is_mapped() function
to check mappedness of ivi_views. The functions checks:
   - the view is on a layer's order list
   - the layer is on a screen
   - the layer and view's ivi_surface are visible

If ivi_view is not mapped but weston_view is still mapped,
we have to unmap the weston_view with weston_view_unmap() call.

Reported-by: Michael Olbrich <m.olbrich@pengutronix.de>
Signed-off-by: Emre Ucan <eucan@de.adit-jv.com>

Move the implementation from commit_screen_list to
build_view_list function

Signed-off-by: Emre Ucan <eucan@de.adit-jv.com>

Remoting plugin support streaming image of virtual output on drm-backend
to remote output. By appending remote-output section in weston.ini,
weston loads remoting plugin module and creates virtual outputs via
remoting plugin. The mode, host, and port properties are configurable in
remote-output section.

This plugin send motion jpeg images to client via RTP using gstreamer.
Client can receive by using following pipeline of gst-launch.

gst-launch-1.0 rtpbin name=rtpbin \
   udpsrc caps="application/x-rtp,media=(string)video,clock-rate=(int)90000,
		encoding-name=JPEG,payload=26" port=[PORTNUMBER] !
          rtpbin.recv_rtp_sink_0 \
   rtpbin. ! rtpjpegdepay ! jpegdec ! autovideosink \
   udpsrc port=[PORTNUMBER+1] ! rtpbin.recv_rtcp_sink_0 \
   rtpbin.send_rtcp_src_0 !
          udpsink port=[PORTNUMBER+2] sync=false async=false

where, PORTNUMBER is specified in weston.ini.

Signed-off-by: Tomohito Esaki <etom@igel.co.jp>

Xwayland block SIGUSR1 signal for handling this signal. However, if some
weston plugins creates additional threads before xwayland is loaded,
this signal get delivered these threads and causes weston quit.
Therefore, we should set up SIGUSR1 blocking early so that these threads
can inherit the setting when created.

Signed-off-by: Tomohito Esaki <etom@igel.co.jp>

Add support virtual output for streaming image to remote output by
remoting-plugin which will be added by the patch:
"Add remoting plugin for output streaming."
The gbm bo of virtual output is the linear format.

Virtual output is implemented based on a patch by Grigory Kletsko
<grigory.kletsko@cogentembedded.com>.

Signed-off-by: Tomohito Esaki <etom@igel.co.jp>

Store usage flags of gbm bo in drm_output in order to specify the bo
format for each output. A following patch will add a new type of
drm_output which requires different gbm_bo_flags.

Signed-off-by: Tomohito Esaki <etom@igel.co.jp>

Add new API to gl-renderer interface for providing fence sync fd. the
backend can wait for GPU rendering by this API.

Signed-off-by: Tomohito Esaki <etom@igel.co.jp>

Add new member to store number of planes to drm_fb structure.

Signed-off-by: Tomohito Esaki <etom@igel.co.jp>

Emil Velikov authored 6 years ago and Emil Velikov committed 6 years ago

In the case where CreateContext/MakeCurrent fails, we still want to
know what the EGL driver is capable of.

Move the EGL info printing, just after the eglInitialize() call to
ensure that.

Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

Emil Velikov authored 6 years ago and Emil Velikov committed 6 years ago

Split the two into separate functions. Former requires an initialized
EGL display, while the latter a current context.

We will use that distinction with the next patch.

Signed-off-by: Emil Velikov <emil.velikov@collabora.com>

We are currently formatting 32-bit signed integers into 8 byte buffers,
which are too small, causing the compiler to complain. Update the buffer
size to the minimum required value of 12 bytes: 1 for the sign, 10 for
the number, and 1 for the null byte terminator.

Signed-off-by: Alexandros Frantzis <alexandros.frantzis@collabora.com>

Not every output will have a backlight control, and even if it does we
may just not be able to find it. Not having backlight control isn't an
error, so don't spam the log with it, as doing so can confuse users into
thinking this is an actual error which is responsible for their real
problems.

Signed-off-by: Daniel Stone <daniels@collabora.com>

Not having swap_buffers_with_damage could cause a performance impact on
some backends, but at least on GBM it causes no issues. It also seems to
confuse users into thinking it's a legitimate error which could explain
session slowness.

Similarly with buffer_age, whilst we do lose a little bit of performance
by not being able to do partial renders, it is not a great deal, and the
user is unlikely to be able to do anything about it in any event.

Remove the warning; we print the full extension list at startup, so we
already have enough information from the logs to easily diagnose any
real errors.

Signed-off-by: Daniel Stone <daniels@collabora.com>