Even though it is part of wayland.xml, wl_shell is not mandatory for
compositors to implement, and may be removed in a future version.
Signed-off-by: Demi Marie Obenour <demiobenour@gmail.com>
FreeBSD doesn't support mremap [1], so we have a fallback
implementation based on munmap+mmap. We memcpy from the old memory
region to the new one, however this is unnecessary because the new
mapping references the same file as the old one.
Use msync to make sure any pending write is flushed to the underlying
file before we map the new region.
[1]: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=59912
Signed-off-by: Simon Ser <contact@emersion.fr>
Commit 5a981ee8 implemented a fallback path for platforms which do not
support mremap() such as FreeBSD.
To do so, the file descriptor for the mmap() is not closed immediately
but instead kept as long as the pool exists.
That induces more file descriptors kept open for longer, which in turn
may cause problems as wl_shm may be using a lot of file descriptors,
especially with Xwayland which can create a lot of pixmaps on behalf of
its X11 clients.
For platforms where mremap() is available, keeping those file
descriptors opened is a bit of a waste and may cause exhaustion of file
descriptors sooner that before commit 5a981ee8.
Only keep the mmap() file descriptor open on platforms which do not
implement mremap() and close it immediately as before on others.
Signed-off-by: Olivier Fourdan <ofourdan@redhat.com>
Closes: https://gitlab.freedesktop.org/xorg/xserver/-/issues/1283
Initialiaze the entire msghdr struct to 0 before use.
Example of the report fixed with this change:
AddressSanitizer:DEADLYSIGNAL
=================================================================
==20035==ERROR: AddressSanitizer: SEGV on unknown address 0x2dad4dbffffa0d (pc 0x0055555c7488 bp 0x007fffffc760 sp 0x007fffffc760 T0)
==20035==The signal is caused by a READ memory access.
#0 0x55555c7488 in read_msghdr(void*, __sanitizer::__sanitizer_msghdr*, long) (/home/ftrvx/w/_/uxn/bin/uxnemu+0x77488)
#1 0x55555c810c in sendmsg (/home/ftrvx/w/_/uxn/bin/uxnemu+0x7810c)
#2 0x7ff7f2db20 in wl_connection_flush /home/ftrvx/q/wayland/build/../src/connection.c:315:10
#3 0x7ff7f2d014 in wl_display_flush /home/ftrvx/q/wayland/build/../src/wayland-client.c:2154:9
#4 0x7ff7e80bc0 (/lib/libSDL2-2.0.so.0+0x104bc0)
#5 0x7ff7e523b0 (/lib/libSDL2-2.0.so.0+0xd63b0)
#6 0x7ff7e534e4 (/lib/libSDL2-2.0.so.0+0xd74e4)
#7 0x7ff7e535e8 (/lib/libSDL2-2.0.so.0+0xd75e8)
#8 0x7ff7daad54 (/lib/libSDL2-2.0.so.0+0x2ed54)
#9 0x7ff7dab130 (/lib/libSDL2-2.0.so.0+0x2f130)
#10 0x555565bb40 in main /home/ftrvx/w/_/uxn/src/uxnemu.c:519:2
#11 0x7ff7f62484 in libc_start_main_stage2 /builddir/musl-1.1.24/src/env/__libc_start_main.c:94:2
AddressSanitizer can not provide additional info.
SUMMARY: AddressSanitizer: SEGV (/home/ftrvx/w/_/uxn/bin/uxnemu+0x77488) in read_msghdr(void*, __sanitizer::__sanitizer_msghdr*, long)
==20035==ABORTING
Signed-off-by: Sigrid Solveig Haflínudóttir <ftrvxmtrx@gmail.com>
Fix invalid usage of bools which violates official meson specification and thus
breaks muon, an implementation of meson written in C.
Signed-off-by: illiliti <illiliti@protonmail.com>
wl_signal_emit doesn't handle well situations where a listener removes
another listener. This can happen in practice: wlroots and Weston [1]
both have private helpers to workaround this defect.
wl_signal_emit can't be fixed without breaking the API. Instead,
introduce a new function. Callers need to make sure to always remove
listeners when they are free'd.
[1]: https://gitlab.freedesktop.org/wayland/weston/-/merge_requests/457
Signed-off-by: Simon Ser <contact@emersion.fr>
Signed-off-by: Alexandros Frantzis <alexandros.frantzis@collabora.com>
These events are for high-resolution wheels. Each logical wheel click
represents an axis value120 value of 120, fractions of a wheel click an
integer fraction thereof. The new event replaces
wl_pointer.axis_discrete for wl_pointer versions 8 and above.
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Not checking the result of wl_array_add() can cause writes past the end of the
allocated buffer if realloc fails.
Signed-off-by: Aleksandr Mezin <mezin.alexander@gmail.com>
Fixes the following warning:
WARNING: add_languages is missing native:, assuming languages are wanted for
both host and build.
Signed-off-by: Simon Ser <contact@emersion.fr>
Fixes the following warning:
WARNING: Project targeting '>= 0.56.0' but tried to use feature deprecated
since '0.55.0': ExternalProgram.path. use ExternalProgram.full_path() instead
Signed-off-by: Simon Ser <contact@emersion.fr>
When multiple threads are used, output from different threads was intermixed in one line. That what breaking parsing of the log messages. Now, intermixing is prevented by using a memstream.
Signed-off-by: Alexander Irion <alexander_irion@mentor.com>
Protocol documentation mistakenly specified that wl_shm.format
events are sent at connection setup time, but that's impossible
(clients do not yet know ID of wl_shm object at the time) and
contradicts implementation in libwayland.
Signed-off-by: Mikhail Gusarov <dottedmag@dottedmag.net>
meson.build_root() returns the parent's build directory if Wayland is
a subproject. This fails with:
error: tag OUTPUT_DIRECTORY: Output directory '<parent-build-dir>/doc/doxygen' does not exist and cannot be created
Instead, use meson.project_build_root(), which returns the subproject's
build directory.
Signed-off-by: Simon Ser <contact@emersion.fr>
XcursorLibraryLoadImages() function is unused and not exported according to
objdump, so its removal should be an ABI compatible change.
Signed-off-by: Vlad Zahorodnii <vlad.zahorodnii@kde.org>
Several tests in this suite use setting and checking client.display_stopped (in
test-compositor.h) to synchronise between threads. This is a data race because
display_stopped is a non-atomic int. Fix this by making it an atomic_bool
instead. We don't need to change the access code because reads and writes are
sequentially consistent by default.
This can be reproduced (with both clang and gcc) by running
```
meson -Db_sanitize=thread build
cd build
ninja
meson test
```
Signed-off-by: Fergus Dall <sidereal@google.com>
When allocating memory for structs, use zalloc instead of malloc.
This ensures the memory is zero-initialized, and reduces the risk
of forgetting to initialize all struct fields.
Signed-off-by: Simon Ser <contact@emersion.fr>
Fixes the following warning:
WARNING: You should add the boolean check kwarg to the run_command call.
It currently defaults to false,
but it will default to true in future releases of meson.
See also: https://github.com/mesonbuild/meson/issues/9300
Signed-off-by: Simon Ser <contact@emersion.fr>
Set explicitly the C standard to use to make sure we don't use
features not available on our target platforms.
Signed-off-by: Simon Ser <contact@emersion.fr>
Currently libwayland assumes GNU extensions will be available, but
doesn't define the C standard to use. Instead, let's unconditionally
enable POSIX extensions, and enable GNU extensions on a case-by-case
basis as needed.
Signed-off-by: Simon Ser <contact@emersion.fr>
tc_client_fd_leaks and tc_client_fd_leaks_exec are currently the exact
same test. It seems clear from the name that the latter was intended to
spawn sanity_fd_leak_exec instead of sanity_fd_leak.
Fixes#121
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
Add a note that the request only changes the size of the memory mapping
and does not touch the mapped file.
Signed-off-by: Max Ihlenfeldt <mihlenfeldt@igalia.com>
The specification left the position and order of file
descriptors unspecified. Specify that
- order of file descriptors is maintained
- position of file descriptors is bounded, but loose
Signed-off-by: Mikhail Gusarov <dottedmag@dottedmag.net>
Make it easier to use Wayland as a Meson subproject by overriding
dependencies we define. This allows to easily build Wayland as a
subproject like so:
subproject('wayland', required: false, default_options: ['documentation=false'])
After this statement, the wayland-* dependencies will use the subproject
instead of the system if available.
Signed-off-by: Simon Ser <contact@emersion.fr>
Clearly specify that switching focus within the same client
doesn't mean a new selection will be send.
Signed-off-by: Jan Grulich <jgrulich@redhat.com>
The four new formats added (all 16 bpc, RGB colorspace) are very useful
for applications providing high bit depth content and rendering their
buffers on CPU, as computations with 16 bit unsigned integers are often
more efficient than with the (half float, 10 bit) alternative formats.
Signed-off-by: Manuel Stoeckl <code@mstoeckl.com>
In 11623e8f, SIGBUS handlers aren't set if F_SEAL_SHRINK is configured on
the memory. This helps avoid setting up handlers with cooperative clients;
however, if an application gives an incorrect size, the compositor would
access it anyways, figuring SIGBUS is impossible, and crash.
This can be fixed by simply removing the seal-checking logic and always
setting the signal handler. However, it seems that fstat can give the size
of the memfd, so we can check that the size we were told is within the
region. Since it's sealed to shrinking, it must never be shrunk in future,
so we can really (hopefully) ignore SIGBUS.
I was worried that fstat wasn't supported for this, but shm_overview(7) does
mention that it is a possible use.
The best solution would likely be avoiding SIGBUS entirely with
MAP_NOSIGBUS, but that hasn't been merged yet and wouldn't help systems
without it (e.g. with older kernels).
A proof-of-concept of this crash is attached with the merge request. Running
it with this patch gives an invalid-shm error, which is correct.
Signed-off-by: Duncan McIntosh <duncan82013@live.ca>
Add a note about pre-multiplied alpha for all wl_shm formats.
Include an escape hatch in the spec to allow other protocol
extensions to override this.
Signed-off-by: Simon Ser <contact@emersion.fr>
Closes: https://gitlab.freedesktop.org/wayland/wayland/-/issues/61
The client side closure traces have incorrect object ids for new server
generated objects. This is because create_proxies() overwrites the id in
'n' type arguments by storing a pointer to the actual object in the 'o'
field of the union.
Getting back to an id from this pointer requires accessing a structure
that isn't visible outside of wayland-client.c.
Add a function pointer to fish the correct value out of the argument and
pass it to wl_closure_print.
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
This is inspired from xdg-output-unstable-v1. This allows clients to
get the name and description without having to use xdg_output. This
should eventually allow us to restrict xdg_output to clients like
Xwayland.
The name is a unique non-persistent user-friendly string that can be
used to refer to an output. This can be used by Wayland clients to
refer to a specific wl_output (e.g. across processes or in CLI
arguments).
The description is a non-unique user-friendly string that can be
displayed to the user.
Signed-off-by: Simon Ser <contact@emersion.fr>
References: https://gitlab.freedesktop.org/wayland/wayland-protocols/-/issues/7
The ABI of a shared library on Linux is given by a major version, which
is part of the SONAME and is incremented (rarely) on incompatible
changes, and a minor version, which is part of the basename of the
regular file to which the SONAME provides a symlink.
Until now, the ABI minor version was hard-coded, which means we can't
tell which of a pair of Wayland libraries is newer (and therefore
likely to have more symbols and/or fewer bugs).
libwayland-egl already had ABI major version 1, so we can use the
"marketing" version number as the ABI major.minor version number
directly, so Wayland 1.19.90 would produce
libwayland-egl.so.1 -> libwayland-egl.so.1.19.90.
libwayland-cursor and libwayland-server have ABI major version 0,
and OS distributions don't like it when there's a SONAME bump for no
good reason, so use their existing ABI major version together with
the "marketing" minor version:
libwayland-cursor.so.0 -> libwayland-cursor.so.0.19.90.
If the Wayland major version number is incremented to 2, we'll have to
rethink this, so add some error() to break the build if/when that
happens. Assuming that Wayland 2.0 would involve breaking changes,
the best way would probably to bump all the SONAMEs to
libwayland-foo.so.2.
Resolves: https://gitlab.freedesktop.org/wayland/wayland/-/issues/175
Signed-off-by: Simon McVittie <smcv@collabora.com>
This is meant to address the issue where the call to 'wl_surface.attach'
is done by e.g. Vulkan, meaning applications cannot affect the values of
the offset passed as the x and y arguments.
The lack of ability to pass these is problematic when using the Vulkan
for e.g. drawing DND surfaces, as the buffer offset is used to implement
the drag icon hotspots.
Closes: https://gitlab.freedesktop.org/wayland/wayland/-/issues/148
Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Reviewed-by: Simon Ser <contact@emersion.fr>
