Scanout assignement of views was previously conditioned by views being
opaque, as to allow views with non-opaque buffers reach overlays.

Allow for non-opaque buffers to reach the scanout as well
(as to avoid the case where the last view will try reaching the overlay
 but can not longer due other views being assigned to them).

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>

Under some circumstances, where the HW allows underlays we should
try to assign the views to them. The restriction no longer hold
if view is not opaque and there could be a plane available for it.

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>

overlays/scanout planes

This basically removes the restriction in which we were forced into renderer
mode of compositing when determining that the views would be occluded if
placed in the overlays. We turn it back on once we exhaused (for the
lack of a better term) all available planes the HW made
available to us.

If zpos is available and could place the view in the plane, we use the
zpos property previously read and set the zpos in the plane state (to
apply). If the end, if we managed to used all of them,
we will re-instantiate the restriction.

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>

We need some way to compute the no of planes that have zpos such that we
can determine when we can no longer ascertain that the HW planes, and,
implicily their zpos property, can be used to assign views to them,
and finally can remove the restriction in place when views are
overlay occluded.

This adds also a function to reset to the initial value such that we can
re-use next time the propose state is executed.

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>

Might prove useful at a later point of time when retrieving the
property.

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>

Functional no change, as nobody makes use of it.

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>

placing non-opaque views into the overlay plane

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>

Currently, a check is missing for the case if the HDCP Content Type
property is requested, but is not supported by the driver.

Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>

Modify weston.ini man page to update that drm-backend supports HDCP.

Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>

In case of enforced protection mode, the renderer takes care of
censoring the protected content when the output recording is going on.
But in case of relaxed protection mode, the client must be notified to
avoid showing the protected content, if the output recording is on.

This patch handles the case, where the content-protection is enabled
with relaxed protection mode, and notifies the client, whenever the
recording is started or stopped.

Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>

Harish Krupo authored Jul 05, 2019 and Ankit Nautiyal committed Aug 26, 2019

Contents on an ouput are captured when screenshooter/recorder/screen
sharing is enabled. In such cases the protected content must
be censored to ensure that it is not recorded along with unprotected
content. This is a required only when the surface protection is in
enforced mode.

Signed-off-by: Harish Krupo <harishkrupo@gmail.com>
Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>

Currently, the idle task for updating surface protection is scheduled
in case of change in the output mask of a surface or in case of change
in protection status of an output.
This patch adds a function for reusing the code to schedule the
idle-tasks, that can be called whenever there is a chance of a change
in the protection status of a surface.

Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>

The member disable_planes of weston_output signifies the recording
status of the output, and is incremented and decremented from various
places. This patch provides helper functions to increment and decrement
the counter. These functions can then be used to do processing, before
and after the recording has started or stopped.

Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>

Currently drm-layer supports HDCP1.4 using connector property:
Content Protection. This property if available for a platform, can be
read and set for requesting content-protection.
Also, the patch series [1] adds HDCP2.2 support in drm, and patch [2]
adds support to send udev events for change in connector properties,
made by the kernel.

This patch adds these HDCP connector properties in weston, and exposes
the content-protection support to the client for drm-backend.

It adds the enums to represent 'Content Protection' and 'Content Type'
connector properties exposed by drm layer. It adds a member
'protection' in drm_output_state, to store the desired protection
from the weston_output in the drm-backend output-repaint cycle. This
is then used to write the HDCP connector properties for the drm_heads
attached to the drm_output.

The kernel sends uevents to the user-space for any change made by it
in the "Content Protection" connector property. No event is sent in
case of change in the property made by the user-space.
It means, when there is a change of the property value from "DESIRED"
to "ENABLE" i.e. successful authentication by the kernel, a uevent
will be generated, but in case of userspace requesting for disabling
the protection by writing "UNDESIRED" into the property, no uevent
will be generated.

This patch also adds support for handling new udev events for HDCP
connector property changes. Any such change, triggers change in the
weston_head's current_protection.

[1] https://patchwork.freedesktop.org/series/57233/#rev7
[2] https://patchwork.freedesktop.org/patch/303903/?series=57233&rev=7



Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>

Daniel Stone authored Aug 19, 2019 and Simon Ser committed Aug 23, 2019

62626cbf ensures that the GL render will not render a view's content
to the screen when the surface has requested a higher content-protection
level than the output currently offers.

When the HDCP MR was split into the core content-protection support in !83
and specific DRM support for HDCP in !48 (not yet landed), this opened a
hole where the DRM backend could promote a view to a hardware plane,
even if the output offered a lower protection level than the surface
wanted to enforce.

In the DRM backend, check the desired protection level, and refuse to
promote the view to a hardware plane if the output does not offer
sufficient protection. This will lead to presentation falling back to
the renderer, which may censor the content, reduce quality, etc.

Signed-off-by: Daniel Stone <daniels@collabora.com>
Fixes: 4b6e73d6 ("libweston: Add support to set content-protection for a weston_surface")

Manuel Stoeckl authored Aug 19, 2019 and Simon Ser committed Aug 23, 2019

This ensures that the default signal action doesn't kill weston-terminal
when the terminal tries to paste into a pipe whose read end has already
been shut down. (For example, a pipe from a misconfigured program or from
one which crashes/exits before the terminal calls write().)

Signed-off-by: Manuel Stoeckl <code@mstoeckl.com>

Otherwise 'log_extensions()' will not know how to properly format the
data.

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
Reported-by: Pekka Paalanen <pekka.paalanen@collabora.com>

Information is needed for 'vlog()' and 'vlog_continue()' (others
depend on them).

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>

Stefan Agner authored Jun 20, 2019 and Simon Ser committed Aug 16, 2019

The function pixman_blt may return false in case there is no
accelerated blit function available. In this case the remote shared
screen stays black.

This has been observed on Weston compiled for aarch64. In currrent
pixman 0.38.4 there is no accelerated pixman_blt function for
aarch64 available.

Use pixman_image_composite32 instead which is guaranteed to have a
working fallback implementation.

Fixes: #253

Signed-off-by: Stefan Agner <stefan@agner.ch>

Stefan Agner authored Jun 20, 2019 and Simon Ser committed Aug 16, 2019

Calculate damage region after resizing the cache image. This
avoids unnecessary calculation of damaged regions on resize,
makes sure that the whole screen is considered damaged on
resize and simplifies error handling.

Signed-off-by: Stefan Agner <stefan@agner.ch>

Stefan Agner authored Jun 20, 2019 and Simon Ser committed Aug 16, 2019

Make sure damage region gets properly unreferenced in error cases.

Signed-off-by: Stefan Agner <stefan@agner.ch>

Stefan Agner authored Jun 16, 2019 and Simon Ser committed Aug 16, 2019

Calculate y_orig separately first makes it easer to understand the
code and aligns with how pixels are read in screenshooter.c.

Signed-off-by: Stefan Agner <stefan@agner.ch>

This was caused by weston_wm_handle_xfixes_selection_notify() calling
weston_seat_set_selection() with a NULL source, apparently only
sometimes when closing an Xwayland window.

Manuel Stoeckl authored Aug 07, 2019 and Simon Ser committed Aug 12, 2019

Check return values for wl_display_dispatch_* functions, so that
the program stops running when the compositor that it is connected
to crashes.

Signed-off-by: Manuel Stoeckl <code@mstoeckl.com>

Marius Vlad authored Aug 06, 2019 and Simon Ser committed Aug 12, 2019

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
Reported-by: Pekka Paalanen <pekka.paalanen@collabora.com>

Fixes the following warn/error when using combination of flags like
building with debug, when disabling optimization and/or when enabling ASAN:

../shared/option-parser.c:61:1: error: control reaches end of non-void function [-Werror=return-type]

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>

No need to duplicate messages.

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>

Simon Ser authored Jul 19, 2019 and Pekka Paalanen committed Aug 02, 2019

A few things have changed:

- Meson is used instead of autotools
- Wayland and Weston releases are not synchronized anymore
- Artifact deployment happens via wayland.freedesktop.org's Git repo

While at it, also convert the file to Markdown. Instructions to locally install
Xwayland/libinput have been dropped.

Signed-off-by: Simon Ser <contact@emersion.fr>

Alexandros Frantzis authored Aug 02, 2019 and Pekka Paalanen committed Aug 02, 2019

Although we already supported minor version 2 of the explicit sync
protocol, we couldn't advertise it previously, since it was not in any
released version of wayland-protocols. With the release of
wayland-protocols 1.18, which includes minor version 2 of this protocol,
and the recent update in weston to require 1.18, we can now safely
advertise minor version 2.

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

Jonas Ådahl authored Jul 25, 2019 and Pekka Paalanen committed Aug 02, 2019

Signed-off-by: Jonas Ådahl <jadahl@gmail.com>

Stefan Agner authored Jun 16, 2019 and Simon Ser committed Aug 01, 2019

Commit 4fc5dd00 ("compositor: add capability CAPTURE_YFLIP")
introduced a capability flag which indicates whether y-flipping is
necessary. As already indicated in that commit message, it seems
that pixman flipps the y-axis only due to historic reasons.

Drop y-flipping and use the WESTON_CAP_CAPTURE_YFLIP flag to
indicate that y-flipping is not necessary. This simplifies code
and improves screen share performance (on my test by about 3% down
to 18% CPU load on the sharing instance of Weston).

Signed-off-by: Stefan Agner <stefan@agner.ch>

Morgane Glidic authored Jul 22, 2019 and Pekka Paalanen committed Jul 31, 2019

To fully allow parallel-installation of libweston, we have to make sure
anything that is implemented in libweston is in a versioned directory.

Signed-off-by: Quentin Glidic <sardemff7+git@sardemff7.net>

When building without GL renderer the compiler prints the following
warning:
  ../clients/subsurfaces.c: In function ‘egl_state_create’:
  ../clients/subsurfaces.c:225:35: warning: passing argument 1 of
  ‘weston_platform_get_egl_display’ makes pointer from integer without a
  cast [-Wint-conversion]
    225 |   weston_platform_get_egl_display(EGL_PLATFORM_WAYLAND_KHR,
        |                                   ^~~~~~~~~~~~~~~~~~~~~~~~
        |                                   |
        |                                   int
...

Define the fallback implementation of weston_platform_get_egl_display
to take an integer which is the underlaying datatype of EGLenum.

Signed-off-by: Stefan Agner <stefan@agner.ch>