This fixes the following leaks for
weston_curtain/weston_buffer_create_solid_rgba when shutting down the
compositor:
#0 0x7f9170372987 in __interceptor_calloc ../../../../src/libsanitizer/asan/asan_malloc_linux.cpp:154
#1 0x7f915bfeb8b7 in zalloc ../include/libweston/zalloc.h:38
#2 0x7f915bfec71d in weston_curtain_create ../shell-utils/shell-utils.c:150
#3 0x7f915bfd9e51 in shell_ensure_fullscreen_black_view ../desktop-shell/shell.c:2082
#4 0x7f915bfda2a9 in shell_configure_fullscreen ../desktop-shell/shell.c:2118
#5 0x7f915bfdc72d in desktop_surface_committed ../desktop-shell/shell.c:2538
#6 0x7f915bfa3ef5 in weston_desktop_api_committed ../libweston-desktop/libweston-desktop.c:159
#7 0x7f915bfae778 in weston_desktop_xdg_toplevel_committed ../libweston-desktop/xdg-shell.c:746
#8 0x7f915bfb0d45 in weston_desktop_xdg_surface_committed ../libweston-desktop/xdg-shell.c:1374
#9 0x7f915bfa7382 in weston_desktop_surface_surface_committed ../libweston-desktop/surface.c:174
#10 0x7f916fe628a6 in wl_signal_emit /home/mvlad/install-amd64/include/wayland-server-core.h:481
#11 0x7f916fe7c0e2 in weston_surface_commit_state ../libweston/compositor.c:4062
#12 0x7f916fe7c161 in weston_surface_commit ../libweston/compositor.c:4068
#13 0x7f916fe7c6ef in surface_commit ../libweston/compositor.c:4146
#14 0x7f916fc847e9 (/usr/lib/x86_64-linux-gnu/libffi.so.8+0x77e9)
#0 0x7f9170372987 in __interceptor_calloc ../../../../src/libsanitizer/asan/asan_malloc_linux.cpp:154
#1 0x7f916fe62aa3 in zalloc ../include/libweston/zalloc.h:38
#2 0x7f916fe7398d in weston_buffer_create_solid_rgba ../libweston/compositor.c:2603
#3 0x7f915bfec879 in weston_curtain_create ../shell-utils/shell-utils.c:162
#4 0x7f915bfd9e51 in shell_ensure_fullscreen_black_view ../desktop-shell/shell.c:2082
#5 0x7f915bfda2a9 in shell_configure_fullscreen ../desktop-shell/shell.c:2118
#6 0x7f915bfdc72d in desktop_surface_committed ../desktop-shell/shell.c:2538
#7 0x7f915bfa3ef5 in weston_desktop_api_committed ../libweston-desktop/libweston-desktop.c:159
#8 0x7f915bfae778 in weston_desktop_xdg_toplevel_committed ../libweston-desktop/xdg-shell.c:746
#9 0x7f915bfb0d45 in weston_desktop_xdg_surface_committed ../libweston-desktop/xdg-shell.c:1374
#10 0x7f915bfa7382 in weston_desktop_surface_surface_committed ../libweston-desktop/surface.c:174
#11 0x7f916fe628a6 in wl_signal_emit /home/mvlad/install-amd64/include/wayland-server-core.h:481
#12 0x7f916fe7c0e2 in weston_surface_commit_state ../libweston/compositor.c:4062
#13 0x7f916fe7c161 in weston_surface_commit ../libweston/compositor.c:4068
#14 0x7f916fe7c6ef in surface_commit ../libweston/compositor.c:4146
#15 0x7f916fc847e9 (/usr/lib/x86_64-linux-gnu/libffi.so.8+0x77e9)
We do not migrate the weston_curtain destruction from
desktop_surface_removed() to desktop_shell_destroy_surface() because
we'd want have the curtain removed before the animation starts.
If we were to move the black view destruction, *and* only handle it from
desktop_shell_destroy_surface() the animation runs but the black curtain
will be removed right at the end, effectively diminishing the effect of
the animations.
To this end, we keep both of the two worlds, if the client terminates on
its own, we keep the same animation effect, but if the compositor is
shutting down we destroy it immediately.
We remove wl_list_for_each_safe() and instead loop each time to avoid
using a stale pointer iterator which could cause a UAF as the shsurf
would be free'ed.
Suggested-by: Pekka Paalanen <pekka.paalanen@collabora.com>
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
This fixes the following weston_view leak at compositor shutdown:
#0 0x7f4250247987 in __interceptor_calloc ../../../../src/libsanitizer/asan/asan_malloc_linux.cpp:154
#1 0x7f424fd37aa3 in zalloc ../include/libweston/zalloc.h:38
#2 0x7f424fd3a05f in weston_view_create ../libweston/compositor.c:386
#3 0x7f423be7be6a in weston_desktop_surface_create_desktop_view ../libweston-desktop/surface.c:364
#4 0x7f423be7c0a8 in weston_desktop_surface_create_view ../libweston-desktop/surface.c:404
#5 0x7f423beae91d in desktop_surface_added ../desktop-shell/shell.c:2273
#6 0x7f423be77db1 in weston_desktop_api_surface_added ../libweston-desktop/libweston-desktop.c:138
#7 0x7f423be80c73 in weston_desktop_xdg_toplevel_ensure_added ../libweston-desktop/xdg-shell.c:362
#8 0x7f423be8207a in weston_desktop_xdg_toplevel_committed ../libweston-desktop/xdg-shell.c:697
#9 0x7f423be84d45 in weston_desktop_xdg_surface_committed ../libweston-desktop/xdg-shell.c:1374
#10 0x7f423be7b382 in weston_desktop_surface_surface_committed ../libweston-desktop/surface.c:174
#11 0x7f424fd378a6 in wl_signal_emit /home/mvlad/install-amd64/include/wayland-server-core.h:481
#12 0x7f424fd510e2 in weston_surface_commit_state ../libweston/compositor.c:4062
#13 0x7f424fd51161 in weston_surface_commit ../libweston/compositor.c:4068
#14 0x7f424fd516ef in surface_commit ../libweston/compositor.c:4146
#15 0x7f424fb597e9 (/usr/lib/x86_64-linux-gnu/libffi.so.8+0x77e9)
With commit 'libweston, desktop-shell: Add a wrapper for weston_surface
reference' we've removed an explicit weston_view_destroy() call due to a
UAF which would've happen if we had close animations enabled, upon
terminating a client. In that patch I've incorrectly wrote this happened
when animations are off, but in fact it happened when they're on, see the
following trace:
READ of size 8 at 0x616000026498 thread T0
#0 0x7f757fba8797 in weston_signal_emit_mutable ../shared/signal.c:52
#1 0x7f757fb4bba1 in weston_view_destroy ../libweston/compositor.c:2269
#2 0x7f756bca89c0 in desktop_shell_destroy_surface ../desktop-shell/shell.c:275
#3 0x7f756bcb379e in fade_out_done_idle_cb ../desktop-shell/shell.c:2228
#4 0x7f757faec1da in wl_event_loop_dispatch_idle ../src/event-loop.c:969
#5 0x7f757faec31d in wl_event_loop_dispatch ../src/event-loop.c:1032
#6 0x7f757faea114 in wl_display_run ../src/wayland-server.c:1408
#7 0x7f757ff777ba in wet_main ../compositor/main.c:3589
#8 0x55f765c8d17d in main ../compositor/executable.c:33
#9 0x7f757fd997fc in __libc_start_main ../csu/libc-start.c:332
#10 0x55f765c8d099 in _start (/home/mvlad/install-amd64/bin/weston+0x1099)
0x616000026498 is located 24 bytes inside of 608-byte region [0x616000026480,0x6160000266e0)
freed by thread T0 here:
#0 0x7f758004c4d7 in __interceptor_free ../../../../src/libsanitizer/asan/asan_malloc_linux.cpp:127
#1 0x7f757fb4bdc8 in weston_view_destroy ../libweston/compositor.c:2295
#2 0x7f757fb4c14d in weston_surface_unref ../libweston/compositor.c:2334
#3 0x7f756bca898b in desktop_shell_destroy_surface ../desktop-shell/shell.c:273
#4 0x7f756bcb379e in fade_out_done_idle_cb ../desktop-shell/shell.c:2228
#5 0x7f757faec1da in wl_event_loop_dispatch_idle ../src/event-loop.c:969
This patch re-introduces it to avoid leaking the view upon compositor
shutdown, but it does it in tandem with weston_desktop_surface_unlink_view(),
(which was added in a later patch) and before weston_surface_unref() call.
This way we should be safe to terminate/close clients with additional views
created by libweston-desktop (pop-ups), but also in other different situations.
Verified it in the following circumstances:
- terminating a client with close animation on
- terminating a client with close animations off
- shutting down the compositor with clients running, with and
without close animations
- terminating top-level clients with additional pop-ups with both with
and without close animations
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
Fix an apparent copy and paste error in resize code. I'm not sure anything
sets the relevant callback that would lead to height being different than
width, so there's no easy way to demonstrate a bug, but this change
appears to rectify the intent of the code.
Reported-by: Hideyuki Nagase <hideyukn@microsoft.com>
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
This documents the fact that other views are handled implictly by
libweston-desktop, and we shouldn't attempt to destroy indiscriminately
views that aren't created by desktop-shell.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
This patch fixes the following trace:
#0 0x7f07d1bcecfa in weston_desktop_surface_get_surface ../libweston-desktop/surface.c:585
#1 0x7f07d1bfc9b8 in move_grab_motion ../desktop-shell/shell.c:1499
#2 0x7f07e539f841 in notify_motion ../libweston/input.c:1794
#3 0x7f07e1e8ace4 in handle_pointer_motion ../libweston/libinput-device.c:132
#4 0x7f07e1e8cad5 in evdev_device_process_event ../libweston/libinput-device.c:535
#5 0x7f07e1e89311 in udev_input_process_event ../libweston/libinput-seat.c:208
#6 0x7f07e1e8932f in process_event ../libweston/libinput-seat.c:218
#7 0x7f07e1e8935f in process_events ../libweston/libinput-seat.c:228
#8 0x7f07e1e8940a in udev_input_dispatch ../libweston/libinput-seat.c:239
#9 0x7f07e1e89437 in libinput_source_dispatch ../libweston/libinput-seat.c:249
#10 0x7f07e53122b1 in wl_event_loop_dispatch ../src/event-loop.c:1027
#11 0x7f07e5310114 in wl_display_run ../src/wayland-server.c:1408
#12 0x7f07e579c7ba in wet_main ../compositor/main.c:3589
#13 0x555611d6b17d in main ../compositor/executable.c:33
#14 0x7f07e55be7fc in __libc_start_main ../csu/libc-start.c:332
#15 0x555611d6b099 in _start (/home/mvlad/install-amd64/bin/weston+0x1099)
A highly unlikely, but still valid operation, is to close/destroy the
window while still having it grabbed and moved around, basically having
an in-flight destruction of grabbed moving window. Another situation
would be that the client terminates abruptly (crashing for instance),
while being dragged which might take down the compositor.
This could happen for both touch/pointer grab operations and could cause
a NULL pointer access while accessing desktop_surface when being used
to retrieve the underlying weston_surface.
With this patch we check for a valid desktop_surface and return early
if that's not the case.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
Moving weston_desktop_surface_unlink_view() to
desktop_shell_destroy_surface() makes sure we don't leak the underlying
weston_desktop_view when tearing/shutting down the compositor.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
Creates a distinct view, separated from the one created by
libweston-desktop, in order to avoid a potential ownership fight with
libweston-desktop upon destroying the view. Upon weston_desktop_surface
destruction, libweston-desktop inflicts damage on the view it creates,
so we need the view to be alive at that time.
This wasn't such an issue before because we had different destruction paths but
with commit 'desktop-shell: Do not leave views in layers upon shell
destruction' all of the destruction paths now land in the same spot
+ handle compositor tear down.
Note as we still use the same weston_surface we'll keep the previous
construct where we were taking a reference to keep it alive.
The original view is destroyed when releasing the ownership, while for
the view created in this patch we handle the destruction directly upon
compositor tear down.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
Similar to how we do it with drm_fb ref counts, increase a reference
count and return the same object.
Plug-in in desktop-shell when we map up the view in order to survive a
weston_surface destruction.
Astute readers will notice that this patch removes weston_view_destroy()
while keeping the balance between removing and adding a
weston_surface_unref() call in desktop_shell_destroy_surface().
The reason is to let weston_surface_unref() handle destruction on its
own. If multiple references are taken, then weston_surface_unref()
doesn't destroy the view, it just decreases the reference, with
a latter call to weston_surface_unref() to determine if the view
should be destroyed as well. This situation happens if we have
close animation enabled, were we have more than one reference taken: one
when mapping the view/surface and when when the surface itself was created,
(what we call, a weak reference).
If only a single reference is taken (for instance if we don't have close
animations enabled) then this weston_surface_unref()
call is inert as that reference is not set-up, leaving libweston to
handle the view destruction.
Following that with a weston_view_destroy() explicit call would cause a
UAF as the view was previous destroyed by a weston_surface_unref() call.
A side-effect of not keeping the weston_view_destroy() call would
happen when tearing down the compositor. If close animations are enabled,
weston_surface_unref() would not destroy the view, and because
weston_view_destroy() no longer exists, we would still have the
view in the other layers by the time we check-up if layers
have views present.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
Make it obvious that weston_surface has a reference counting happening
and destruction of the weston_surface happens when the last
weston_surface reference has been accounted for.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
For a surface that is already fullscreen making it maximized means to
exit fullscreen then set to it maximized. Instead of doing it, refuse to
do anything until the user explicitly performs that operation.
With this approach we follow other DE (desktop environments) which would
not perform any operation until the user exits fullscreen state.
Fixes#321
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
Exposay was done as a showcase for what we could do with Weston and an
efficient compositing pipeline. This was mostly with the old
vendor-specific Raspberry Pi backend which could actually process that
many surfaces bypassing the GPU.
Given enough bitrot, Exposay is now pretty exemplary as what _not_ to do
in a Weston shell - particularly the way it manipulates existing
weston_views rather than create its own non-destructive stack.
As it's annoying technical debt, a terrible example to others, and not a
very compelling showcase in 2022, just delete it.
Signed-off-by: Daniel Stone <daniels@collabora.com>
This is a minor re-work of how we de-activate and activate the surfaces
in desktop-shell. As activate() is being used for handling keyboard
input events, this basically rectifies that such that activation
happens only if the passed surface is different from the currently
focused surface.
Fixes: #593
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
By some dark magic this accidentally doesn't crash because there are
zeroes in the right places at startup, but black_surface_get_label()
very much expects surface_private to be a view.
Let's hand craft a new bespoke label function for this use.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
shell-utils contains a number of helpers which are currently in use by
both desktop-shell and kiosk-shell. In order to extend this use to
fullscreen-shell as well (which can benefit from reusing the
weston_curtain infrastructure to be able to create solid-colour views
which may or may not be opaque, as well as one function within
fullscreen-shell which was copied wholesale to shell-utils), we need to
create a separate Meson dependency object, and avoid the existing
pattern of including the source from shared/ within the source list for
each shell.
This requires creating a new top-level directory for these shared helper
functions which are required by each shell, but are not part of
libweston in and of itself.
shell-utils depends on libweston-desktop; libweston-desktop depends on
libweston; libweston depends on shared.
Thus it is not possible to expose a dependency object from the shared/
directory which declares a dependency on the libweston-desktop
dependency, as Meson processes directories in order and resolves
variable references as they are parsed.
In order to break this deadlock, this commit creates a new top-level
directory called 'shell-utils' containing only this file, which can be
parsed by Meson after libweston-desktop (making the libweston-desktop
Meson dependency variable available to the build file to declare a
dependency on that), but before the shells (making the new Meson
depenendency object available to each shell which wishes to use it).
This commit contains no functional changes to any observable code.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Use the common infrastructure we have, rather than open-coding again.
In changing to the common weston_curtain infrastructure which was
introduced in a previous commit, there are two small functional
derivations.
After adding the view to a layer, we explicitly call
weston_view_geometry_dirty(). This is believed to have no functional
impact, as weston_views have their geometry-dirty flag set when they are
created.
weston_surface_damage() is also called to ensure that the surface is
scheduled for a repaint. As there is currently no good common mechanic
in the common code to ensure that output repaint will occur, due to the
damage propagating outwards from the weston_surface, we are forced to
call this to predictably ensure that the output will be repainted after
we have made this change to the scene graph which should result in the
user observing the new content being repainted.
It is possible that these changes are not strictly required in order for
the correct behaviour to occur. However, it is felt that explicitly
adding these calls is perhaps the least fragile way to ensure that the
behaviour continues to be correct and breakage is not observed,
especially with both view mapping and surface damage identified as areas
for future work which could be beneficial to Weston. If it is felt that
these calls can be removed, then this is certainly possible at a later
stage.
Lastly, there are many users within desktop-shell of the common pattern
of creating a weston_curtain and inserting it into the scene graph to be
painted. These users have been an area of both theoretical concern and
real-life observed fragility and breakage recently. By making each user
follow a common and predictable pattern of usage, each user is no longer
its own special case. This should make it easier to make the
desktop-shell codebase more robust, not only simplifying the codebase,
but improving observed behaviour.
In doing this, it is hoped that future structural and interface changes
become easier to make, allowing us to improve some of the more difficult
corners of the libweston API.
Signed-off-by: Daniel Stone <daniels@collabora.com>
This will allow us to create a solid weston_buffer as well, since we
need to store that separately.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Just as we do for fullscreen backgrounds, reuse the curtain infrastructure
for focus animations.
This introduces a small functional change, in that the surface's output
is no longer directly assigned. Instead, we call
weston_view_set_output() ourselves. As setting the weston_view's
position (done from the common helper function of weston_curtain_create
which has been introduced in previous commits) will call
weston_view_set_position(), the view's geometry will be dirtied as a
result.
When the geometry of a weston_view is dirty, it is marked to be updated,
which will occur whilst traversing the view list during output repaint.
This occurs for every view which is currently assigned to a layer; when
building the view list, any view reachable through the view list whose
geometry is dirty will have its position recalculated and an output
assigned. Doing so results in the surface's current output being
updated.
It is believed that there is no functional impact from the
weston_surface not having a primary output assigned between creation and
output repaint being called.
Signed-off-by: Daniel Stone <daniels@collabora.com>
desktop-shell's focus surfaces want to reuse this, but they don't want
to capture the input, instead allowing it to fall through.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Rationalise it a little bit so we don't need pre-declarations of static
functions, and the order of creation more closely matches the others,
including making the same calls to explicitly set the output.
Doing this makes the behaviour match the other users of the same code
pattern. In making them the same, we make desktop-shell code a little
less magically divergent where people might wonder what the correct
pattern is to use. After we have moved all users to a uniform pattern,
later commits are then able to migrate these users to common helper
code, which reduces code duplication, improves code clarity as it is no
longer necessary to wonder about the exact semantics of every
special-case user of this common pattern, and makes it easier to make
interface changes which improve and clarify the patterns which are
prevalent throughout the desktop-shell code.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Dirtying the geometry only sets a flag on the view saying that the
geometry is dirty, so we don't need to do it twice back-to-back.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Opaque regions are in surface co-ordinate space, not global co-ordinate
space, so the region should be anchored to (0,0).
Signed-off-by: Daniel Stone <daniels@collabora.com>
Not all solid-colour views want to be opaque: sometimes we use them with
non-opaque alpha values in order to shade views underneath them.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Given that we have a struct for argument params, we might as well use it
rather than have them split between the struct and native params. For
consistency between the implementations, this also includes a shift from
float to int positioning for the base offset within the compositor's
global co-ordinate space.
Signed-off-by: Daniel Stone <daniels@collabora.com>
The name implied that it was a surface in and of itself, rather than
parameters used by a helper to create a surface and view.
Rename it now that we have weston_curtain as a name, and clean up
initialisers.
Signed-off-by: Daniel Stone <daniels@collabora.com>
create_solid_color_surface actually returns a weston_view that it
creates internally. Since weston_solid_color_view is long and dull,
rename it to weston_curtain.
Signed-off-by: Daniel Stone <daniels@collabora.com>
desktop_shell_removed() won't get called when we shut down, so
explicitly destroy the fullscreen black view.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Zoom is a neat trick, but in its current form it's very hard to test
and maintain.
It also causes output damage to scale outside of the output's boundaries,
which leads to an extra clipping step that's only necessary when zoom
is enabled.
Remove it to simplify desktop-shell and compositor.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
This properly takes care of removing any of the views being added into
different layers upon shell destruction. Aggregates the shell_surface
destruction into one place to make things much more clearer. Keep the
animation part only the weston_desktop_surface destruction rather than
doing it at shutdown.
Fixes#509.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
Starting with commit 'desktop-shell: Embed keyboard focus handle code
when activating' we are iterating over all the seats when removing a
weston_desktop_surface to be able to invalidate the activate surface.
Upon compositor tear down the shell destroy might invalidate seats
before removal of the weston_desktop_surface making the shell_seat
itself invalid. This wasn't apparent at that time because we're not
handling at that the removal of surfaces from layers.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
Attempting to perform a switch on a surface (already) closed will trip
the assert in activate(), so check if we have a weston_desktop_surface
before trying to activate it.
Fixes#543
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
We no longer make use of the keyboard_focus_listener so remove it
entirely.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
Suggested-by: Derek Foreman <derek.foreman@collabora.com>
We shouldn't be constrained by having a keyboard plugged-in, so avoid
activating/de-activating the window/surface in the keyboard focus
handler and embed it straight into the window activation part.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
The shsurf is calloc'ed so the surface count is always 0. Not only
that but the surface is not set as active by default, so there's no
need to de-activate it.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
When the surface type of input panel is set as an overlay panel, it's
expected to be shown at near the input cursor. But the current
implementation shows it at center-bottom on the desktop at first then
move it to the correct position while typing.
Signed-off-by: Takuro Ashie <ashie@clear-code.com>
The keyboard focus listener, caps changed and pointer focus listener
were missing when destroying the seat. These are necessary to avoid
using weston_desktop object even if it was destroyed, which happens due
to a focus out event and ultimately handled by the keyboard focus notify
callback.
Once the seats are destroyed (and implictly the focus handlers) we're
safe to destroy weston_desktop object as well.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
This properly tracks the seats and with it, destroys them before
destroying weston_desktop object.
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
Fixes an issue where subsurface extending outside of the main surface
wasn't damaged when minimized resulting in left-over content.
Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Follow-up from commit 'desktop-shell: don't run fade animation if
compositor is inactive' where the reference was dropped directly,
instead of using weston_surface_destroy().
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
When a window is closed, Weston will, by default, run a fade out animation and
defer destroying the underlying surface until it completes. However, if the
compositor is sleeping, and therefore not rendering any frames, this animation
will *never* complete. Therefore, if windows are repeatedly created and
destroyed while in sleep mode, these surfaces will keep accumulating, and since
the buffers attached to them may be backed by an fd, eventually the ulimit will
be reached resulting in a potential crash or other errors.
This can be demonstrated repeatedly launching and killing an X11 application
with Xwayland running.
while true; do xterm & pid=$!; sleep 0.5; kill $pid; done
As soon as the compositor goes to sleep, one can observe a steadily growing
list of dmabufs in the output of lsof.
As a fix, desktop_surface_removed should check whether the compositor is active
before kicking off the fade animation. If it is not, it should instead drop the
extra reference taken in desktop_surface_committed and then destroy the surface
immediately.
Signed-off-by: Erik Kurzinger <ekurzinger@nvidia.com>
This fixes the following ASan detected leaks:
Direct leak of 88 byte(s) in 1 object(s) allocated from:
#0 0x7f8c3455f330 in __interceptor_malloc (/lib/x86_64-linux-gnu/libasan.so.5+0xe9330)
#1 0x7f8c33c60906 in wl_event_loop_add_timer ../../git/wayland/src/event-loop.c:571
#2 0x7f8c2ff98f46 in shell_fade_init ../../git/weston/desktop-shell/shell.c:4211
#3 0x7f8c2ff9e1da in wet_shell_init ../../git/weston/desktop-shell/shell.c:5266
#4 0x7f8c3443ede5 in wet_load_shell ../../git/weston/compositor/main.c:956
#5 0x7f8c3444fdb9 in wet_main ../../git/weston/compositor/main.c:3434
#6 0x55878ad3bfc6 in execute_compositor ../../git/weston/tests/weston-test-fixture-compositor.c:432
#7 0x55878ad3f9fa in weston_test_harness_execute_as_client ../../git/weston/tests/weston-test-runner.c:528
#8 0x55878ad2fbd6 in fixture_setup ../../git/weston/tests/viewporter-test.c:46
#9 0x55878ad2fc58 in fixture_setup_run_ ../../git/weston/tests/viewporter-test.c:48
#10 0x55878ad3ffaf in main ../../git/weston/tests/weston-test-runner.c:661
#11 0x7f8c340b409a in __libc_start_main ../csu/libc-start.c:308
#12 0x55878ad2f769 in _start (/home/pq/build/weston-meson/tests/test-viewporter+0xc769)
Indirect leak of 856 byte(s) in 1 object(s) allocated from:
#0 0x7f8c3455f518 in calloc (/lib/x86_64-linux-gnu/libasan.so.5+0xe9518)
#1 0x7f8c33c99b73 in zalloc ../../git/weston/include/libweston/zalloc.h:38
#2 0x7f8c33c9cfb1 in weston_surface_create ../../git/weston/libweston/compositor.c:574
#3 0x7f8c2ff98230 in shell_fade_create_surface_for_output ../../git/weston/desktop-shell/shell.c:4059
#4 0x7f8c2ff98df6 in shell_fade_init ../../git/weston/desktop-shell/shell.c:4202
#5 0x7f8c2ff9e1da in wet_shell_init ../../git/weston/desktop-shell/shell.c:5266
#6 0x7f8c3443ede5 in wet_load_shell ../../git/weston/compositor/main.c:956
#7 0x7f8c3444fdb9 in wet_main ../../git/weston/compositor/main.c:3434
#8 0x55878ad3bfc6 in execute_compositor ../../git/weston/tests/weston-test-fixture-compositor.c:432
#9 0x55878ad3f9fa in weston_test_harness_execute_as_client ../../git/weston/tests/weston-test-runner.c:528
#10 0x55878ad2fbd6 in fixture_setup ../../git/weston/tests/viewporter-test.c:46
#11 0x55878ad2fc58 in fixture_setup_run_ ../../git/weston/tests/viewporter-test.c:48
#12 0x55878ad3ffaf in main ../../git/weston/tests/weston-test-runner.c:661
#13 0x7f8c340b409a in __libc_start_main ../csu/libc-start.c:308
#14 0x55878ad2f769 in _start (/home/pq/build/weston-meson/tests/test-viewporter+0xc769)
Indirect leak of 608 byte(s) in 1 object(s) allocated from:
#0 0x7f8c3455f518 in calloc (/lib/x86_64-linux-gnu/libasan.so.5+0xe9518)
#1 0x7f8c33c99b73 in zalloc ../../git/weston/include/libweston/zalloc.h:38
#2 0x7f8c33c9bed5 in weston_view_create ../../git/weston/libweston/compositor.c:365
#3 0x7f8c2ff98251 in shell_fade_create_surface_for_output ../../git/weston/desktop-shell/shell.c:4063
#4 0x7f8c2ff98df6 in shell_fade_init ../../git/weston/desktop-shell/shell.c:4202
#5 0x7f8c2ff9e1da in wet_shell_init ../../git/weston/desktop-shell/shell.c:5266
#6 0x7f8c3443ede5 in wet_load_shell ../../git/weston/compositor/main.c:956
#7 0x7f8c3444fdb9 in wet_main ../../git/weston/compositor/main.c:3434
#8 0x55878ad3bfc6 in execute_compositor ../../git/weston/tests/weston-test-fixture-compositor.c:432
#9 0x55878ad3f9fa in weston_test_harness_execute_as_client ../../git/weston/tests/weston-test-runner.c:528
#10 0x55878ad2fbd6 in fixture_setup ../../git/weston/tests/viewporter-test.c:46
#11 0x55878ad2fc58 in fixture_setup_run_ ../../git/weston/tests/viewporter-test.c:48
#12 0x55878ad3ffaf in main ../../git/weston/tests/weston-test-runner.c:661
#13 0x7f8c340b409a in __libc_start_main ../csu/libc-start.c:308
#14 0x55878ad2f769 in _start (/home/pq/build/weston-meson/tests/test-viewporter+0xc769)
They were found with:
ASAN_OPTIONS=fast_unwind_on_malloc=0,malloc_context_size=50 \
LSAN_OPTIONS=suppressions=/home/pq/git/weston/.gitlab-ci/leak-sanitizer.supp \
./tests/test-viewporter test_viewporter_double_create
Turns out shell_destroy() had an open-coded and outdated copy of the
tear-down sequence, so fixing the leaks in only handle_output_destroy()
was not enough.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
Most other places call a variable like this 'shell_output', so let's do
that here too. The old name was really confusing.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
Daniel Stone
Marius Vlad
Derek Foreman
Takuro Ashie
Jonas Ådahl
Emmanuel Gil Peyrot
Erik Kurzinger
Pekka Paalanen