The gbm-format configuration option can now be specified per-output as
well as in the core config section. If it is not specified it will
default to the format specified in the core section. The
EGL_MESA_configless_context extension is required for this to work. If
this extension is available it will create a context without an
EGLConfig and then it will potentially use a different EGLConfig for
each output.
The gl-renderer interface has been changed so that it takes the EGL
attributes and visual ID in the create_output function as well as in
the create function.
Part of the gl_renderer_setup function only deals with checking EGL
extensions and doesn't need to have a current context. This patch
moves these checks so that they are done during gl_renderer_create
instead of waiting until we have an output. We will need this in a
later patch because some of the EGL extensions will affect how we
create the EGLSurface.
Commit fa1b3055 introducted a regression in the pixman renderer. In
particular, it would not draw properly with any output transform other
than normal, 180, 180-flipped, or 270-flipped. This patch fixes this
regression. The weston-scalar program appears to work normally at all
rotations with this patch. Therefore, this patch should fix the regression
while maintaining the added functionality from fa1b3055.
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
The noop-renderer doesn't read buffer contents, which means bad
buffers go undetected. Thus, read the buffer contents just for
the purpose of triggering SIGBUS (and having the client killed).
Fixes the bad-buffer test when run against the headless backend.
Signed-off-by: Emilio Pozuelo Monfort <emilio.pozuelo@collabora.co.uk>
This lets the compositor know the size of the surface as calculated
in weston_surface_set_size_from_buffer(), and fixes a couple of
tests when using the headless backend.
Signed-off-by: Emilio Pozuelo Monfort <emilio.pozuelo@collabora.co.uk>
Fixes a segfault when using compositor-headless for the test suite
as many tests assume there are input devices and try to use them
through the wl_test interface.
Signed-off-by: Emilio Pozuelo Monfort <emilio.pozuelo@collabora.co.uk>
The client needs to know that the pointer is at a different position in
its surface. We can't send motion as that corresponds to the pointer
actually moving. Leaving the surface and entering at the new position
is a better semantic match and doesn't correspond to pointer motion
or user input.
https://bugs.freedesktop.org/show_bug.cgi?id=71927
This was always a little iffy. At least it could have been a signal,
but we now have focus signal, so lets just use that. We lose
the ability to detect unresponsive clients at key event time, but we
could add that back by adding a key_signal.
eglCreateContext fails with every EGLConfig that
nvidia blob 334.16 provides causing NULL pointer
dereference in gl_renderer_destroy when destroying
fragment and fan bindings.
https://bugs.freedesktop.org/show_bug.cgi?id=74699
Signed-off-by: Mariusz Ceier <mceier+wayland@gmail.com>
Remove the listener for output destroy from weston_view and instead
iterate views owned by the shell in its own output destroy listener.
This simplifies the code a bit since keeping the view listening for the
destroy on the right output was a bit complicated. This also removes the
function pointer output_destroyed from weston_view. The only user for it
was desktop shell, but now this is all handled in shell.c.
Since that signal is per output, it is necessary to track in which
output a view is in so that the signal is handled properly.
Instead, add a compositor wide output moved signal, that is handled by
the shell. The shell iterates over the layers it owns to move views
appropriately.
The input initialization code assumes the outputs have already
been initialized; thus create the outputs first. This fixes a
segfault upon startup. It is also what the drm and fbdev backends
do.
We don't want to send events if the binding is going to handle the touch
event. Also, this restricts touch bindings to only trigger on touch down.
For gesture bindings we want something similar to the motion signal we
have for the pointer.
When we send the pointer motion event, the transform from compositor to
surface coordinates doesn't depend on the resource. Transform the
coordinates up front instead of everytime we send to a resource.
If only the mode or the owner are wrong, do not say both are wrong.
Change the text to state that there's a problem and the current
values, and let the user figure it out.
Signed-off-by: Guillem Jover <guillem@hadrons.org>
The pointer seat->keyboard was set before some possible error returns.
That pointer was left unchanged in case of failure, pointing to an
uninitialized keyboard struct (that was also leaked). If a client sent
a wl_seat::get_keyboard request, that would cause Weston to crash.
Fix this by setting the seat->keyboard pointer only after the keymap
initialization is done and there is no more possibilities for failure.
Also plug the memory leaks on the error path.
https://bugs.freedesktop.org/show_bug.cgi?id=74035
The input region of the cursor surface is set to empty in
pointer_cursor_surface_configure(). Since during the commit process
this function is called before the pending input region is made
current, it empties surface->pending.input instead of surface->input.
But pointer_cursor_surface_configure() is also called from
pointer_set_cursor() in order to map the cursor even if there isn't a
subsequent attach and commit to the cursor surface. In that case,
surface->input is never emptied, since the configure function emptied
only the pending input region and there wasn't a commit that made it
effective.
Fix this by emptying both pending and current input regions. The latter
shouldn't cause problems since the surface can't have a role prior to
being assigned the cursor role, so it shouldn't be mapped in the first
place.
Also change toytoolkit so that it triggers the bug.
https://bugs.freedesktop.org/show_bug.cgi?id=73711
Not doing this would leave a invalid list item in the view's destroy
signal listener list if destroying a seat that had previously lost
keyboard focus.
Signed-off-by: Jonas Ådahl <jadahl@gmail.com>
Adding this comment to explain the behavior:
This macro may not do what you expect. Weston doesn't guarantee any
stable API between 1.X and 1.Y, and thus this macro will return
FALSE on any WESTON_VERSION_AT_LEAST(1,X,0) if the actualy version
is 1.Y.0 and X !=Y). In particular, it fail if X < Y, that is,
1.3.0 is considered to not be "at least" 1.4.0.
If you want to test for the version number being 1.3.0 or above or
maybe in a range (eg 1.2.0 to 1.4.0), just use the WESTON_VERSION_*
defines above directly.
Version number testing is the one thing we can't break in the weston API,
so we'll have to settle for documenting the behavior and recommending
using the version number macros directly.
https://bugs.freedesktop.org/show_bug.cgi?id=74023
If we VT switch away between picking a cursor surface and actually doing
the pageflip in drm_output_repaint(), we never set output->cursor_view to
NULL. Then we unplug all the input devices and as the last pointer device
goes away we destroy the cursor surface. Then when we switch back, we
call drm_output_set_cursor() with an invalid surface and crashes.
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=73566
Use the STAMP_SPACE to make the output mode logging
a little nicer looking.
Signed-off-by: U. Artie Eoff <ullysses.a.eoff@intel.com>
Reviewed-by: Bryce Harrington <b.harrington@samsung.com>
The API to use remoteFx encoding has changed between master and stable 1.1
branch. This patch should fix compilation for both.
This new version adds checks for the freerdp/version.h file
Assigning a string constant (i.e. memory that we didn't allocate)
to a char* pointer and then freeing that pointer is bad news.
Signed-off-by: U. Artie Eoff <ullysses.a.eoff@intel.com>
Handle the case where localtime fails (NULL) and print
something else to indicate localtime is erroneous.
Signed-off-by: U. Artie Eoff <ullysses.a.eoff@intel.com>
The opaque region is in surface coordinates, which we compare to the
output region, which is in compositor coordinates. For non-primary
outputs, that means that the output region is not located at 0,0 but
something like 1920,0 instead. That means that the output region isn't
contained in the surface opaque region and then we decide we can't scan
out from it.
Instead, compare the surface opaque region to the output region
translated to 0,0.
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=7348i5
This fixes a regression caused by either 918f2dd4 or da75ee1d. In
particular, if a client called commit without attaching a buffer and if the
compositor had already released its reference to the buffer, then a size of
0x0 would be set on the surface. In particular, this affects the wayland
backend because it frequently sends only a frame request in order to cause
a refresh.
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
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
Jonas Ådahl
Neil Roberts
Kristian Høgsberg
Jason Ekstrand
Emilio Pozuelo Monfort
Mariusz Ceier
Jasper St. Pierre
Ander Conselvan de Oliveira
Quentin Glidic
Guillem Jover
U. Artie Eoff
Hardening
Bryce W. Harrington