If we destroy the text_input (e.g. due to surface deactivation) whilst
the IM manager (the client holding the input_method_context resource,
usually weston-keyboard) sends an event, we can hit a segfault in
text-backend. This is because we free the text_input structure, but
don't actually clear the context->input structure even when we send the
deactivate event.
This is clearly intended to be catered for, since context->input is
always checked for NULL before we relay any events.
This is enough to fix one cause of text-test failing, but it's
ultimately error-prone until we have a no-op test-shell; there is still
a race where weston-desktop-shell can launch weston-keyboard before
text-test manages to bind zwp_text_input_manager.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Maniphest Tasks: https://phabricator.freedesktop.org/T7615
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
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 config can now be retrieved with a new function defined in weston.h,
wet_get_config(weston_compositor*).
Signed-off-by: Giulio Camuffo <giuliocamuffo@gmail.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-by: Quentin Glidic <sardemff7+git@sardemff7.net>
They belong in the compositor rather than libweston since they
set signals handlers, and a library should not do that behind its
user's back. Besides, they were using functions in main.c already
so they were not usable by other compositors.
Signed-off-by: Giulio Camuffo <giuliocamuffo@gmail.com>
Reviewed-by: Quentin Glidic <sardemff7+git@sardemff7.net>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Reviewed-by: Mariusz Ceier <mceier+wayland@gmail.com>
Reviewed-by: Jan Arne Petersen <janarne@gmail.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Reviewed-by: Mariusz Ceier <mceier+wayland@gmail.com>
Reviewed-by: Jan Arne Petersen <janarne@gmail.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
We really don't want to deal with sighups and pids. It's far easier
to just deal with the client destroyed signal to respawn the input method.
Inspiration taken from Pekka's commit 826dc14ec4
This fixes a potential crash on shutdown. If the input method client
pointer is still set (sighup handler hasn't run) but the client is
already destroyed, a call to text_backend_destroy() will try to destroy
the client a second time resulting in a segfault.
Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
The binding is only used for preventing launch, and we've already got the
pid for that.
With multiple seats there are multiple bindings, so if the most recent
binding unbinds it will clear the pointer, so using it like this is risky
anyway.
Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
Previously we tried to launch the input method every time a seat was
created, and the launch function would notice it was running and not
bother to actually launch.
Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Keyboards and pointers aren't freed when devices are removed, so we should
really be testing keyboard_device_count and pointer_device_count in most
cases, not the actual pointers. Otherwise we end up with different
behaviour after removing a device than we had before it was inserted.
This commit renames the touch/keyboard/pointer pointers and adds helper
functions to get them that hide this complexity and return NULL when
*_device_count is 0.
Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Jonas Ådahl <jadahl@gmail.com>
We used to rely on the order in which the
weston_compositor::destroy_signal callbacks happened, to not access
freed memory. Don't know when, but this broke at least with ivi-shell,
which caused crashes in random places on compositor shutdown.
Valgrind found the following:
Invalid write of size 8
at 0xC2EDC69: unbind_input_panel (input-panel-ivi.c:340)
by 0x4E3B6BB: destroy_resource (wayland-server.c:537)
by 0x4E3E085: for_each_helper.isra.0 (wayland-util.c:359)
by 0x4E3E60D: wl_map_for_each (wayland-util.c:365)
by 0x4E3BEC7: wl_client_destroy (wayland-server.c:675)
by 0x4182F2: text_backend_notifier_destroy (text-backend.c:1047)
by 0x4084FB: wl_signal_emit (wayland-server-core.h:264)
by 0x4084FB: main (compositor.c:5465)
Address 0x67ea360 is 208 bytes inside a block of size 232 free'd
at 0x4C2A6BC: free (vg_replace_malloc.c:473)
by 0x4084FB: wl_signal_emit (wayland-server-core.h:264)
by 0x4084FB: main (compositor.c:5465)
Invalid write of size 8
at 0x4E3E0D7: wl_list_remove (wayland-util.c:57)
by 0xC2EDEE9: destroy_input_panel_surface (input-panel-ivi.c:191)
by 0x4E3B6BB: destroy_resource (wayland-server.c:537)
by 0x4E3BC7B: wl_resource_destroy (wayland-server.c:550)
by 0x40DB8B: wl_signal_emit (wayland-server-core.h:264)
by 0x40DB8B: weston_surface_destroy (compositor.c:1883)
by 0x40DB8B: weston_surface_destroy (compositor.c:1873)
by 0x4E3B6BB: destroy_resource (wayland-server.c:537)
by 0x4E3E085: for_each_helper.isra.0 (wayland-util.c:359)
by 0x4E3E60D: wl_map_for_each (wayland-util.c:365)
by 0x4E3BEC7: wl_client_destroy (wayland-server.c:675)
by 0x4182F2: text_backend_notifier_destroy (text-backend.c:1047)
by 0x4084FB: wl_signal_emit (wayland-server-core.h:264)
by 0x4084FB: main (compositor.c:5465)
Address 0x67ea370 is 224 bytes inside a block of size 232 free'd
at 0x4C2A6BC: free (vg_replace_malloc.c:473)
by 0x4084FB: wl_signal_emit (wayland-server-core.h:264)
by 0x4084FB: main (compositor.c:5465)
Invalid write of size 8
at 0x4E3E0E7: wl_list_remove (wayland-util.c:58)
by 0xC2EDEE9: destroy_input_panel_surface (input-panel-ivi.c:191)
by 0x4E3B6BB: destroy_resource (wayland-server.c:537)
by 0x4E3BC7B: wl_resource_destroy (wayland-server.c:550)
by 0x40DB8B: wl_signal_emit (wayland-server-core.h:264)
by 0x40DB8B: weston_surface_destroy (compositor.c:1883)
by 0x40DB8B: weston_surface_destroy (compositor.c:1873)
by 0x4E3B6BB: destroy_resource (wayland-server.c:537)
by 0x4E3E085: for_each_helper.isra.0 (wayland-util.c:359)
by 0x4E3E60D: wl_map_for_each (wayland-util.c:365)
by 0x4E3BEC7: wl_client_destroy (wayland-server.c:675)
by 0x4182F2: text_backend_notifier_destroy (text-backend.c:1047)
by 0x4084FB: wl_signal_emit (wayland-server-core.h:264)
by 0x4084FB: main (compositor.c:5465)
Address 0x67ea368 is 216 bytes inside a block of size 232 free'd
at 0x4C2A6BC: free (vg_replace_malloc.c:473)
by 0x4084FB: wl_signal_emit (wayland-server-core.h:264)
by 0x4084FB: main (compositor.c:5465)
Looking at the first of these, unbind_input_panel() gets called when the
text-backend destroys its helper client which has bound to input_panel
interface. This happens after the shell's destroy_signal callback has
been called, so the shell has already been freed.
The other two errors come from
wl_list_remove(&input_panel_surface->link);
which has gone stale when the shell was destroyed
(shell->input_panel.surfaces list).
Rather than creating even more destroy listeners and hooking them up in
spaghetti, modify text-backend to not hook up to the compositor destroy
signal. Instead, make it the text_backend_init() callers' responsibility
to also call text_backend_destroy() appropriately, before the shell goes
away.
This fixed all the above Valgrind errors, and avoid a crash with
ivi-shell when exiting Weston.
Also using desktop-shell exhibited similar Valgrind errors which are
fixed by this patch, but those didn't happen to cause any crashes AFAIK.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
Reviewed-By: Derek Foreman <derekf@osg.samsung.com>
Removed duplicate definitions of the container_of() macro and
refactored sources to use the single implementation.
Signed-off-by: Jon A. Cruz <jonc@osg.samsung.com>
Reviewed-by: Peter Hutterer <peter.hutterer@who-t.net>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
This allows a user to explicitly disable the input
method by setting path to blank;
Signed-off-by: Murray Calavera <murray.calavera@gmail.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
a following patch will be moving text init call into shell
modules, which will be called much later than in current code
Signed-off-by: Murray Calavera <murray.calavera@gmail.com>
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
commit 78d00e45cc renamed text_model to text_input
This cleans up remaining uses of the word "model"
Reviewed-by: Jan Arne Petersen <janarne@gmail.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Multi-seat configurations currently break the text-backend, crashing
weston. This is an attempt to clean up any crashes and have somewhat
sensible input panel behavior with multi-seat.
Store a link to the manager that created a text_input, use this to
ensure that only a single panel gets popped up at a time, since there
is only one manager.
Replace deactivate_text_input with deactivate_input_method: multiple
input methods may focus the same text_input, so deactivating a text_input
is weird in multi-seat and confusing to perform.
In destroy_input_method_context set the context's input_method's context
pointer to NULL to prevent a dangling pointer.
Reviewed-by: Jan Arne Petersen <janarne@gmail.com>
Reviewed-by: Bryce Harrington <bryce@osg.samsung.com>
Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
A keyboard might not be present in a seat, so check that before
dereferencing keyboard related pointers.
Also, use the keyboard pointer we set to shorten the code a little bit.
Signed-off-by: Derek Foreman <derekf@osg.samsung.com>
Reviewed-by: Jan Arne Petersen <janarne@gmail.com>
Fix recently introduced compiler warnings:
desktop-shell/shell.c: In function 'shell_configuration':
desktop-shell/shell.c:588:10: warning: ignoring return value of
'asprintf', declared with attribute warn_unused_result [-Wunused-result]
src/screenshooter.c: In function ‘screenshooter_binding’:
src/screenshooter.c:291:10: warning: ignoring return value of
‘asprintf’, declared with attribute warn_unused_result [-Wunused-result]
src/text-backend.c: In function ‘text_backend_configuration’:
src/text-backend.c:944:10: warning: ignoring return value of ‘asprintf’,
declared with attribute warn_unused_result [-Wunused-result]
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
desktop shell and weston keyboard both refer to themselves prefixed by
LIBEXECDIR, however this is only valid once installed. make check will
currently either fail or run pre-existing versions.
This patch adds a way to override that location by setting the env var
WESTON_BUILD_DIR - which is then set by the test env script so make check
will test the versions in the build directory regardless of whether they're
installed or not.
Reviewed-by: Pekka Paalanen <pekka.paalanen@collabora.co.uk>
An error makes the client exit, which cleans up the resources anyway.
Note (Jason Ekstrand):
This is safe for two reasons. First, we should be handling object
destruction nicely anyway. Second, in each of these cases, the resources
don't have any implementation or destruction set so it has absolutely no
effect on the rest of weston whether we destroy it now or later.
There's a small window between the input method (eg the on-sreen keyboard)
receiving the deactivate and destroying the context, where the keyboard may
send requests, which we forward to the destroyed input method. Fix this
by setting the contexts model to NULL right away and then avoid sending
events if context->model is NULL.
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=69490
A grab can potentially allocate memory and would normally end the grab
itself, freeing the allocated memory in the process. However at in some
situations the compositor may want to abort a grab. The grab owner still
needs to free some memory and abort the grab properly. To do this a new
function 'cancel' is introduced in all the grab interfaces instructing
the grabs owner to abort the grab.
This patch also hooks up grab cancelling to seat device releasing and
when the compositor looses focus, which would potentially leak memory
before.
Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Hi Kristian,
Here's a new patch for ref counting weston_xkb_info, as suggested.
So a seat created with a NULL keymap will now point to the global xkb_info.
Prior to freeing the memory in which the link node for the signal is
emedded we should remove the link node from the list to prevent the list
from being corrupted.
https://bugs.freedesktop.org/show_bug.cgi?id=67231
This commit sets the version numbers for all added/created objects. The
wl_compositor.create_surface implementation was altered to create a surface
with the same version as the underlying wl_compositor. Since no other
"child interfaces" have version greater than 1, they were all hard-coded to
version 1.
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
With the change to move free()ing of the wl_resource into wayland-server, we now have
a few cases where we double free the resource in the destructor. This patch
removes those.
This is the first in what will be a series of weston patches to convert
instances of wl_resource to pointers so we can make wl_resource opaque.
This patch handles weston_surface and should be the most invasive of the
entire series. I am sending this one out ahead of the rest for review.
Specifically, my machine is not set up to build XWayland so I have no
ability to test it fully. Could someone please test with XWayland and let
me know if this causes problems?
Because a surface may be created from XWayland, the resource may not always
exist. Therefore, a destroy signal was added to weston_surface and
everything used to listen to surface->resource.destroy_signal now listens
to surface->destroy_signal.
AC_USE_SYSTEM_EXTENSIONS enables _XOPEN_SOURCE, _GNU_SOURCE and similar
macros to expose the largest extent of functionality supported by the
underlying system. This is required since these macros are often
limiting rather than merely additive, e.g. _XOPEN_SOURCE will actually
on some systems hide declarations which are not part of the X/Open spec.
Since this goes into config.h rather than the command line, ensure all
source is consistently including config.h before anything else,
including system libraries. This doesn't need to be guarded by a
HAVE_CONFIG_H ifdef, which was only ever a hangover from the X.Org
modular transition.
Signed-off-by: Daniel Stone <daniel@fooishbar.org>
[pq: rebased and converted more files]
Just the same as it is done in shell.c, if the input method process exits
for any reason, we relaunch it automatically, as it is not possible to
launch a standalone application outside of the weston process.
In v2:
- Proper error message when giving up.
Signed-off-by: Eduardo Lima (Etrunko) <eduardo.lima@intel.com>
This set of changes adds support for searching for a given config file
in the directories listed in $XDG_CONFIG_DIRS if it wasn't found in
$XDG_CONFIG_HOME or ~/.config. This allows packages to install custom
config files in /etc/xdg/weston, for example, thus allowing them to
avoid dealing with home directories.
To avoid a TOCTOU race the config file is actually open()ed during the
search. Its file descriptor is returned and stored in the compositor
for later use when performing subsequent config file parses.
Signed-off-by: Ossama Othman <ossama.othman@intel.com>
struct weston_surface is now the only surface type we have (in core, shell.c
has shell_surface, of course). A lot of code gets simpler and we never
have to try to guess whether an API takes a wl_surface or a weston_surface.
Daniel Stone
Jussi Kukkonen
Pekka Paalanen
Giulio Camuffo
Jonas Ådahl
Derek Foreman
Murray Calavera
Jon Cruz
Bryce Harrington
Jasper St. Pierre
Kristian Høgsberg
Andrew Wedgbury
Rob Bradford
Jason Ekstrand
Eduardo Lima (Etrunko)
Daniel Stone
Eduardo Lima (Etrunko)
Ossama Othman