Found by Address sanitizer on test-devices:
==10640==ERROR: AddressSanitizer: heap-use-after-free on address 0x60c000000580 at pc 0x7fa0f050dcd1 bp 0x7fff41c908e0 sp 0x7fff41c908d8
WRITE of size 8 at 0x60c000000580 thread T0
#0 0x7fa0f050dcd0 in unbind_input_method ../../git/weston/compositor/text-backend.c:852
#1 0x7fa0efd1b20d in destroy_resource ../../git/wayland/src/wayland-server.c:724
#2 0x7fa0efd1f7f1 in for_each_helper ../../git/wayland/src/wayland-util.c:372
#3 0x7fa0efd1fcde in wl_map_for_each ../../git/wayland/src/wayland-util.c:385
#4 0x7fa0efd1b35c in wl_client_destroy ../../git/wayland/src/wayland-server.c:883
#5 0x7fa0f050ea82 in text_backend_destroy ../../git/weston/compositor/text-backend.c:1067
#6 0x7fa0ebb69f2f in shell_destroy ../../git/weston/desktop-shell/shell.c:5012
#7 0x7fa0efd55933 in wl_signal_emit /home/pq/local/include/wayland-server-core.h:478
#8 0x7fa0efd7d061 in weston_compositor_destroy ../../git/weston/libweston/compositor.c:7896
#9 0x7fa0f050a349 in wet_main ../../git/weston/compositor/main.c:3493
#10 0x559c1e794354 in execute_compositor ../../git/weston/tests/weston-test-fixture-compositor.c:432
#11 0x559c1e797dc0 in weston_test_harness_execute_as_client ../../git/weston/tests/weston-test-runner.c:528
#12 0x559c1e786ab8 in fixture_setup ../../git/weston/tests/devices-test.c:39
#13 0x559c1e786b3a in fixture_setup_run_ ../../git/weston/tests/devices-test.c:41
#14 0x559c1e798375 in main ../../git/weston/tests/weston-test-runner.c:661
#15 0x7fa0f016e09a in __libc_start_main ../csu/libc-start.c:308
#16 0x559c1e786769 in _start (/home/pq/build/weston-meson/tests/test-devices+0xc769)
0x60c000000580 is located 0 bytes inside of 120-byte region [0x60c000000580,0x60c0000005f8)
freed by thread T0 here:
#0 0x7fa0f0618fb0 in __interceptor_free (/lib/x86_64-linux-gnu/libasan.so.5+0xe8fb0)
#1 0x7fa0f050df1d in input_method_notifier_destroy ../../git/weston/compositor/text-backend.c:902
#2 0x7fa0efd86d77 in wl_signal_emit /home/pq/local/include/wayland-server-core.h:478
#3 0x7fa0efd98086 in weston_seat_release ../../git/weston/libweston/input.c:3475
#4 0x7fa0ebb0d002 in test_seat_release ../../git/weston/tests/weston-test.c:132
#5 0x7fa0ebb0e197 in device_release ../../git/weston/tests/weston-test.c:314
#6 0x7fa0efca88ed in ffi_call_unix64 (/lib/x86_64-linux-gnu/libffi.so.6+0x68ed)
previously allocated by thread T0 here:
#0 0x7fa0f0619518 in calloc (/lib/x86_64-linux-gnu/libasan.so.5+0xe9518)
#1 0x7fa0f050a8bf in zalloc ../../git/weston/include/libweston/zalloc.h:38
#2 0x7fa0f050e6f1 in text_backend_seat_created ../../git/weston/compositor/text-backend.c:1011
#3 0x7fa0f050e947 in handle_seat_created ../../git/weston/compositor/text-backend.c:1040
#4 0x7fa0efd86d77 in wl_signal_emit /home/pq/local/include/wayland-server-core.h:478
#5 0x7fa0efd97d57 in weston_seat_init ../../git/weston/libweston/input.c:3440
#6 0x7fa0ebb0ce4b in test_seat_init ../../git/weston/tests/weston-test.c:110
#7 0x7fa0ebb0f699 in wet_module_init ../../git/weston/tests/weston-test.c:592
#8 0x7fa0f04f8d69 in wet_load_module ../../git/weston/compositor/main.c:941
#9 0x7fa0f04f914d in load_modules ../../git/weston/compositor/main.c:1012
#10 0x7fa0f0509ec1 in wet_main ../../git/weston/compositor/main.c:3441
#11 0x559c1e794354 in execute_compositor ../../git/weston/tests/weston-test-fixture-compositor.c:432
#12 0x559c1e797dc0 in weston_test_harness_execute_as_client ../../git/weston/tests/weston-test-runner.c:528
#13 0x559c1e786ab8 in fixture_setup ../../git/weston/tests/devices-test.c:39
#14 0x559c1e786b3a in fixture_setup_run_ ../../git/weston/tests/devices-test.c:41
#15 0x559c1e798375 in main ../../git/weston/tests/weston-test-runner.c:661
#16 0x7fa0f016e09a in __libc_start_main ../csu/libc-start.c:308
Fix UAF by resetting wl_resource user data, and ensuring it is valid
before used.
Setting seat->input_method to NULL may not be necessary since it is
being called from seat destroy listener, but added just in case.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
Since commit "libweston: add required_capabilities test suite quirk"
a new function that depends on test_data is being called in wet_main().
We should check if test_data is NULL before calling it, otherwise
we have a segfault when running outside the test suite.
Signed-off-by: Leandro Ribeiro <leandro.ribeiro@collabora.com>
This allows tests to skip when required capabilities are not present.
The output damage test for the shadow buffer case needs this.
required_capabilities is added to struct weston_testsuite_quirks which
is libweston public API just because there is no better place currently.
This is a little weird because the code to check it is in compositor,
not libweston.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
The capability strings will be printed also in another occasion, where
the colon does not fit with the capability description.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
This adds an output section option use-renderer-shadow in weston.ini.
This option is only recognized with headless and DRM backends, because
it requires GL-renderer and does not support resizing outputs.
The option is called use-renderer-shadow because this is what it does
right now. In the future the same setting will be used to turn on more
complex image processing when operational pieces required for color
management land. Once color management is implemented, this option is
expected to be removed. This option allows developer testing of features
to be used to implement color management.
This is a rewrite of "weston.ini: introduce use-shadow-fbo in output
config" by Harish Krupo. The main.c code is structured differently, the
weston.ini option is renamed, and the man page paragraph is moved to
weston.ini.man with different content.
Cc: Harish Krupo <harishkrupo@gmail.com>
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
This adds the libweston capability bit for "color operations" which
refers to a renderer's support for operations needed for color
management. GL-renderer will grow the support while Pixman-renderer will
not, which is why the cap is needed.
To make an example use of the cap, this also adds new API:
weston_output_set_renderer_shadow_buffer(). This is a temporary API to
enable future experimental features. The first such feature will be the
renderer internal shadow buffer, the boolean variable for it taken from
Harish Krupo's "weston.ini: introduce use-shadow-fbo in output config".
Obviously this patch does not implement the renderer shadow buffer. No
renderer sets WESTON_CAP_COLOR_OPS yet so trying to enable it will fail.
The documentation here is deliberately vague, because the bits needed
for color management will come in trickling for a long time until we can
call it color management in any sense. Until then, the temporary API
shall remain, perhaps poorly named.
Cc: Harish Krupo <harishkrupo@gmail.com>
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
shm buffer should be freed in case of `out_pixman_error`, and should
be inserted into shared_output just before the success return.
Signed-off-by: xndcn <xndchn@gmail.com>
The buffer created for flight recorder, the allocations done in 'logger' and
'flight_rec' subscribers were not freed when wl_display_create() fails. So
move them to out_display
Signed-off-by: Anurup M <anurup.mokkil@gmail.com>
Without this patch, the DRM-backend would rewrite the 'require-input',
core section option given by the user.
This removes 'continue_without_input' DRM-backend option and takes into
consideration the cmd line option only if that was passed (Pekka Paalanen).
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
Until now we had the test quirks initialization in wet_main(),
just after calling weston_compositor_create(). But there are
some cases that require the quirks during struct weston_compositor
creation time.
Move test quirks initialization to weston_compositor_create()
in order to cover more use cases for the test quirks mechanism.
Signed-off-by: Leandro Ribeiro <leandro.ribeiro@collabora.com>
We have two functions with the name weston_screenshooter_shoot():
1. screenshot protocol function that the screenshooter
client uses to request screenshots to the compositor
2. libweston function used by the compositor to take
screenshots as requested by the screenshooter client
Until now we had no problem with that, but in the next commits
we are going to use the screenshot protocol in the test suite,
which is also user of libweston. So rename screenshot protocol
function to weston_screenshooter_take_shot() to avoid the conflict.
For consistency, also rename screenshooter_shoot() to
screenshooter_take_shot() in compositor/weston-screenshooter.c
Signed-off-by: Leandro Ribeiro <leandro.ribeiro@collabora.com>
Until now we had struct wet_testsuite_data as an opaque
struct that should be defined by the testsuite of libweston
users. Instead, keep the data as a void * and document that
users are responsible for defining the data type.
Signed-off-by: Leandro Ribeiro <leandro.ribeiro@collabora.com>
There are some specific cases in which we need Weston to
behave differently when running in the test suite. This
adds a new API to allow the tests to select these behaviors.
For instance, in the DRM backend we plan to add a writeback
connector screenshooter. In case it fails for some
reason, it should fallback to the renderer screenshooter
that all other backends use. But if we add a test to
ensure the correctness of the writeback screenshooter,
we don't want it to fallback to the renderer one, we
want it to fail. With this new API we can choose to
disable the fallback behavior specifically for this test.
Signed-off-by: Leandro Ribeiro <leandro.ribeiro@collabora.com>
This was forgotten in "weston: remove SEGV and ABRT handlers". It is
unused.
Fixes: bb707dc0fe
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
This commit alters the way that Weston picks a Wayland display socket
name. Instead of using wl_display_add_socket_auto to look for the first
available name in wayland-0, wayland-1, .... to wayland-32, the code now
checks names wayland-1, wayland-2, .... up to wayland-32.
This change is a workaround for a suboptimal behavior of
libwayland-client. If a client program calls wl_display_connect(NULL) and
the WAYLAND_DISPLAY environment variable is not set, then the program will
by default try to connect to 'wayland-0'. This is a problem when a
computer has a running Wayland compositor but is being accessed in some
other fashion, such as through an X session on a different virtual
terminal, over ssh, etc. Client programs launched through those means may
attempt to connect to an unrelated compositor. Changing libwayland
behavior to remove the default would also work, but a) libraries have
stronger backward compatibility expectations b) that would likely break
more people's setups than just changing Weston would.
Signed-off-by: Manuel Stoeckl <code@mstoeckl.com>
directory
Weston is also a user of the plug-ins, so make use of it. With this
change we unconditionally install the plug-in headers even though
libweston might not be built with support for them.
Suggested-by: Pekka Paalanen <pekka.paalanen@collabora.com>
Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
The missing build dependency was added. The override to disable this
check can be removed.
Signed-off-by: Guillaume Champagne <champagne.guillaume.c@gmail.com>
cms-colord uses cms-helper functions which require lcms2. Therefore,
lcms2 must be added as a build dependency.
Signed-off-by: Guillaume Champagne <champagne.guillaume.c@gmail.com>
In the test suite we may want to run a DRM-backend test on a
non-default seat, which may not have a input device associated.
Weston's default behavior is to not open if input devices are
not found, as it may cause troubles. For instance, Weston can
open but if no input device is set than the user can not
interact or leave it.
Add flag --continue-without-input to DRM-backend so we can run
these types of tests with no input. Notice that this won't force
the compositor to skip opening a input device if it finds it on
the non-default seat.
Signed-off-by: Leandro Ribeiro <leandro.ribeiro@collabora.com>
If users ask explicitly to log to a file, it makes sense to quit
when we fail opening that file. Continuing execution would mean
wasting users' time if they expect to find the log file at the
end of the session.
Signed-off-by: Antonio Caggiano <antonio.caggiano@collabora.com>
Suggested-by: Pekka Paalanen <pekka.paalanen@collabora.com>
When failing to open the log file nothing is reported to the user,
therefore we print a message on stderr when that happens.
Signed-off-by: Antonio Caggiano <antonio.caggiano@collabora.com>
Suggested-by: Pekka Paalanen <pekka.paalanen@collabora.com>
As in aaf35586f4, we want to fail when we are passed an invalid
transform name, not just blindly configure on using the normal
transform. The previous commit missed the callsite from the headless
backend's command-line parsing.
Fix this so that headless fails when an invalid transform is specified
on the command line.
Signed-off-by: Daniel Stone <daniels@collabora.com>
If the output only has a single weston_head attached to it, take its
declared transform as the default transform.
With the previous patches, this allows a device declaring the KMS 'panel
orientation' property (e.g. through DeviceTree) to autoconfigure to the
correct display rotation when running Weston.
Signed-off-by: Daniel Stone <daniels@collabora.com>
If an invalid transformation is provided for an output, fail the output
configuration rather than continuing on using whatever we chose as the
default transform.
After !383, this will result in configurations using the old definition
failing and exiting, rather than continuing on the wrong way around.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Regardless of the default transform passed in, weston_parse_transform
would always return 'normal' if there was an output section. This is
because, if a section was declared for that output, it would ask
weston_config for the transform, with the default being 'normal'.
Fix it so we return the passed-in default transform when we have a
matching output section without a transform key. If the transform is
declared but invalid, we can remove the line resetting to the default
transform, because we've already set the default transform up top.
Signed-off-by: Daniel Stone <daniels@collabora.com>
The test suite wants to start using different output scales, and this is the
easiest API to configure it.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
It was discovered in issue #99 that the implementations of the 90 and 270
degree rotations were actually the inverse of what the Wayland specification
spelled out. This patch fixes the libweston implementation to follow the
specification.
As a result, the behaviour of the the weston.ini transform key also changes. To
force all users to re-think their configuration, the transform key values are
also changed. Since Weston and libweston change their behaviour, the handling
of clients' buffer transform changes too.
All the functions had their 90/270 cases simply swapped, probably due to
confusion of whether WL_OUTPUT_TRANSFORM_* refers to rotating the monitor or
the content.
Hint: a key to understanding weston_matrix_rotate_xy(m, c, s) is that the
rotation matrix is formed as
c -s
s c
that is, it's column-major. This fooled me at first.
Fixing window.c fixes weston-terminal and weston-transformed.
In simple-damage, window_get_transformed_ball() is fixed to follow the proper
transform definitions, but the fix to the viewport path in redraw() is purely
mechanical. The viewport path looks broken to me in the presence of any
transform, but it is not this patch's job to fix it.
Screen-share fix just repeats the general code fix pattern, I did not even try
to understand that bit.
https://gitlab.freedesktop.org/wayland/weston/issues/99
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
Before commit "weston-log: destroy subscriptions with
destruction of subscribers", we had to destroy subscribers
before the log context. Currently there's no required order,
both are valid.
But since we've created log context before the subscribers,
we can destroy it after them. This is a style change and
also a prove that now this order is valid as well.
Signed-off-by: Leandro Ribeiro <leandrohr@riseup.net>
Log subscriber API is not type-safe. File and flight recorder
subscribers are created with functions that return
weston_log_subscriber objects. But there's a problem: to destroy
these objects you have to call the right function for each type
of subscriber, and a user calling the wrong destroy function
wouldn't get a warning.
Merge functions that destroy different types of subscribers, making
the log subscriber API type-safe.
Signed-off-by: Leandro Ribeiro <leandrohr@riseup.net>
The cms-static, desktop-shell, hmi-controller, ivi-shell, and screen-share
modules use symbols from libexec_weston, e.g.:
$ ldd /usr/lib/x86_64-linux-gnu/weston/desktop-shell.so | grep "not found"
libexec_weston.so.0 => not found
Loading these modules from weston happens to work because the weston executable
itself links against libexec_weston, and has an rpath set. Still, these modules
depend on a library in a non-standard location. Adding an rpath could help
static checkers to make sure all shared objects' dependencies are present.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
`no_outputs` is declared on the stack and left uninitialized if no
weston option changing its value is provided.
Signed-off-by: Guillaume Champagne <champagne.guillaume.c@gmail.com>
Allow to set and get one opaque pointer. The test suite will then use this to
pass data from the test runner to the test plugin, so that the plugin can then
run what it needs to. This is useful when the test runner calls wet_main()
directly instead of forking a compositor.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
The only reason why we have both weston_compositor_tear_down() and
weston_compositor_destroy() is that the only we had to destroy
the log context was keeping weston_compositor alive and calling
weston_log_ctx_compositor_destroy().
After commit "weston-log: replace weston_log_ctx_compositor_destroy()
by weston_log_ctx_destroy()", it's not necessary to keep a zombie
weston_compositor just to be able to call
weston_log_ctx_compositor_destroy().
Fold weston_compositor_tear_down() into weston_compositor_destroy(),
as this split is useless now.
Signed-off-by: Leandro Ribeiro <leandrohr@riseup.net>
The function weston_log_ctx_compositor_destroy(), which destroys struct
weston_log_context, takes weston_compositor as argument. We may have a
weston_log_context unlinked from a weston_compositor and currently there
is no way to destroy it.
Add function weston_log_ctx_destroy(), what makes the destruction of
weston_log_context independent of weston_compositor.
With this change, one could destroy a weston_compositor and keep the
related weston_log_context (since now weston_log_context can be destroyed
without the need of a weston_compositor). But if weston_compositor gets
destroyed it's also necessary to destroy weston_log_context::global,
as the debug protocol depends on the compositor. So a listener has been
added to the destroy signal of weston_compositor.
Signed-off-by: Leandro Ribeiro <leandrohr@riseup.net>
Since weston_log_ctx_compositor_create() does not have any relation
with weston_compositor, rename it to weston_log_ctx_create().
Signed-off-by: Leandro Ribeiro <leandrohr@riseup.net>
There's a function named weston_compositor_log_scope_destroy()
but it doesn't take a struct weston_compositor argument.
Rename it to weston_log_scope_destroy(), as the argument is a
struct weston_log_scope.
Signed-off-by: Leandro Ribeiro <leandrohr@riseup.net>
There's a function named weston_compositor_add_log_scope()
but it doesn't take a struct weston_compositor argument.
Rename it to weston_log_ctx_add_log_scope(), as
the log_scope is being added to a log_context.
Also, bump libweston_major to 9.
Signed-off-by: Leandro Ribeiro <leandrohr@riseup.net>
Define the void pointer as pixman_region32_t instead of using it
directly.
Void pointers are used by generic API's so each user can define
a data type as needed. Not doing this is dangerous, since void
pointers implicitly cast from/to any pointer without giving
warnings. So we must define its type instead of using it directly
in a function call.
Signed-off-by: Leandro Ribeiro <leandrohr@riseup.net>
When option-parser is confronted with a boolean option, have it write a
bool rather than treating the value as a pointer to an int32.
(lib)weston already heavily uses bool types internally, so this has the
nice side effect of eliminating quite a few local variables which only
existed as integer shadows of internal boolean variables.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Wayland innovated a lot of cool things, but non-binary boolean values is
the great advances of our time.
Make config_parser_get_bool() work on boolean values, and switch all its
users.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Instead of getting previous_damage from the weston_output struct, get it from
the frame_signal data argument. This will make possible to remove
previous_damage from weston_output after we decide what to do with
output->previous_damage usage in DRM backend.
Signed-off-by: Leandro Ribeiro <leandrohr@riseup.net>
This is necessary for the test harness to be able to execute the compositor
multiple times in the same process. As we never unload opened modules, the
first compositor iteration will leave them all loaded and following compositor
iterations will then have them already loaded.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
This introduces a new convention of checking through the compositor destroy
listener if the plugin is already initialized. If the plugin is already
initialized, then the plugin entry function succeeds as a no-op. This makes it
safe to load the same plugin multiple times in a running compositor.
Currently module loading functions return failure if a plugin is already
loaded, but that will change in the future. Therefore we need this other method
of ensuring we do not double-initialize a plugin which would lead to list
corruptions the very least.
All plugins are converted to use the new helper, except:
- those that do not have a destroy listener already, and
- hmi-controller which does the same open-coded as the common code pattern
did not fit there.
Plugins should always have a compositor destroy listener registered since they
very least allocate a struct to hold their data. Hence omissions are
highlighted in code.
Backends do not need this because weston_compositor_load_backend() already
protects against double-init. GL-renderer does not export a standard module
init function so cannot be initialized the usual way and therefore is not
vulnerable to double-init.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
This was forgetting to remove the compositor destroy listener, which would lead
to use-after-free on compositor tear-down especially on cms-colord init failure
paths. Found by inspection.
Signed-off-by: Pekka Paalanen <pekka.paalanen@collabora.com>
