Re: [PATCH v2 2/6] shell: register output-destroy_signal handler

2013-10-24 Thread Kristian Høgsberg
On Wed, Oct 23, 2013 at 01:58:32PM +0800, Xiong Zhang wrote: setup_output_destroy_handler() deal with output created at drm backend initialize time. handle_output_create() deal with output created by hot plug handler output_destroy_handler is removed when output was unplugged or shell is

Re: Buffer release events (was: Add support for eglSwapInterval)

2013-10-24 Thread Neil Roberts
Hi Thanks for the interesting insights. It seems to me as if the default should always be to just send the event. I think I would vote for leaving the default as it is, ie, queuing the release events. It's really quite a corner case that delaying events has any effect on an application

Re: Buffer release events (was: Add support for eglSwapInterval)

2013-10-24 Thread Axel Davy
On Wed, 23 Oct 2013, Jason Ekstrand wrote: There may also be a way that we can sidestep the whole issue.  (I suggested this to Axel Davy [mannerov] and it worked for him in his wlglamor DDX.)  The solution is to send a wl_display.sync request immediately after the commit.  This will force

Misc. fixes for the rpi backend

2013-10-24 Thread Tomeu Vizoso
Except for the first one, all are intended to fix the RPi backend after the landing of the views work. Most EGL clients won't work yet because of wl_buffer.release events being always queued, though. ___ wayland-devel mailing list

[PATCH 1/5] rpi: Link with EGL if ENABLE_EGL

2013-10-24 Thread Tomeu Vizoso
--- src/Makefile.am | 6 ++ 1 file changed, 6 insertions(+) diff --git a/src/Makefile.am b/src/Makefile.am index 4224495..9925129 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -224,6 +224,12 @@ rpi_backend_la_SOURCES = \ evdev.c

[PATCH 2/5] rpi: Initialize surface's list of views

2013-10-24 Thread Tomeu Vizoso
--- src/rpi-renderer.c | 1 + 1 file changed, 1 insertion(+) diff --git a/src/rpi-renderer.c b/src/rpi-renderer.c index ea48b08..0b99a60 100644 --- a/src/rpi-renderer.c +++ b/src/rpi-renderer.c @@ -371,6 +371,7 @@ rpir_surface_create(struct rpi_renderer *renderer) if (!surface)

[PATCH 4/5] rpi: EGL surfaces need to be swapped always that we have an incoming back buffer

2013-10-24 Thread Tomeu Vizoso
--- src/rpi-renderer.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/src/rpi-renderer.c b/src/rpi-renderer.c index 2db619c..6478838 100644 --- a/src/rpi-renderer.c +++ b/src/rpi-renderer.c @@ -1255,7 +1255,8 @@ rpi_renderer_repaint_output(struct weston_output *base,

[PATCH 3/5] rpi: Fix logging of buffer swaps for the EGL case

2013-10-24 Thread Tomeu Vizoso
--- src/rpi-renderer.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/src/rpi-renderer.c b/src/rpi-renderer.c index 0b99a60..2db619c 100644 --- a/src/rpi-renderer.c +++ b/src/rpi-renderer.c @@ -969,13 +969,14 @@ rpir_surface_swap_pointers(struct rpir_surface *surface)

[PATCH 5/5] rpi: Only check for prematurely destroyed wl_buffers in the EGL case

2013-10-24 Thread Tomeu Vizoso
--- src/rpi-renderer.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/src/rpi-renderer.c b/src/rpi-renderer.c index 6478838..812e6e7 100644 --- a/src/rpi-renderer.c +++ b/src/rpi-renderer.c @@ -1260,8 +1260,9 @@ rpi_renderer_repaint_output(struct weston_output *base,

[PATCH] input: Don't leak the initial keymap

2013-10-24 Thread Rui Matos
weston_xkb_info_create() takes ownership of the xkb_keymap instance so we should drop our reference or we would leak it later if the keymap was changed. --- src/input.c | 1 + 1 file changed, 1 insertion(+) diff --git a/src/input.c b/src/input.c index 2fed718..da89b47 100644 --- a/src/input.c

Re: [RFC 1/5] Add a fullscreen shell protocol

2013-10-24 Thread Jonas Ådahl
On Wed, Oct 23, 2013 at 06:08:29PM -0500, Jason Ekstrand wrote: Jonas, Thanks for the review! On Wed, Oct 23, 2013 at 3:56 PM, Jonas Ådahl jad...@gmail.com wrote: Hi, Using this protocol, how would the fullscreen shell client turn on and off outputs? I can see how clone, next-to and

Re: Misc. fixes for the rpi backend

2013-10-24 Thread Jason Ekstrand
This whole series looks good to me. Thanks Tomeu. When I rebased on your egl patches I couldn't test due to lack of packages in raspbian. --Jason Ekstrand On Oct 24, 2013 9:22 AM, Tomeu Vizoso to...@tomeuvizoso.net wrote: Except for the first one, all are intended to fix the RPi backend after

Re: Buffer release events (was: Add support for eglSwapInterval)

2013-10-24 Thread Jonas Ådahl
On Thu, Oct 24, 2013 at 11:26:08AM +0100, Neil Roberts wrote: Hi Thanks for the interesting insights. It seems to me as if the default should always be to just send the event. I think I would vote for leaving the default as it is, ie, queuing the release events. It's really quite a

Re: input handlig in separate thread

2013-10-24 Thread Eugen Friedrich
Unfortunately the proposal with a separate queue does not work for the use case. i added the wl_seat object to is own event queue but dispatching this queue blocks the egl rendering currently we are still using wayland 1.05 so my hope is that with the wayland 1.3 and wl_display_prepare_read and

Re: input handlig in separate thread

2013-10-24 Thread Kristian Høgsberg
Yes, this won't work property unless you have wayland 1.2. If you have a EGL driver that uses it's own queue, all you need to do is while (!quit) wl_display_dispatch(display); Otherwise, create a queue and move the wl_registry object to it so that you get all events from all other objects

[PATCH 0/6] DirectFB integration

2013-10-24 Thread Denis Oliver Kropp
Hi, this is an updated patch set. Hopefully in the right shape this time, otherwise please let me know. Please see http://www.directfb.org/docs/DirectFB_Foreseeing_2013-10-07.pdf and http://www.directfb.org/docs/DirectFB_EGL_2013-10-07.pdf for details and plans. With this version you can run

[PATCH 1/6] DirectFB integration

2013-10-24 Thread Denis Oliver Kropp
gl-renderer: Add protection against multiple includes of header. --- src/gl-renderer.h | 8 +++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/src/gl-renderer.h b/src/gl-renderer.h index 0342134..1f41c9f 100644 --- a/src/gl-renderer.h +++ b/src/gl-renderer.h @@ -20,6 +20,9 @@

[PATCH 2/6] DirectFB integration

2013-10-24 Thread Denis Oliver Kropp
surface: Add compositor_state to weston_surface for the compositor to keep a per surface context, like renderer does. --- src/compositor.h | 2 ++ 1 file changed, 2 insertions(+) diff --git a/src/compositor.h b/src/compositor.h index 73722b5..2f3b7d0 100644 --- a/src/compositor.h +++

[PATCH 3/6] DirectFB integration

2013-10-24 Thread Denis Oliver Kropp
configure: Add C++ compiler support and enable c++0x --- configure.ac | 3 +++ 1 file changed, 3 insertions(+) diff --git a/configure.ac b/configure.ac index 80a5d69..30fb4a3 100644 --- a/configure.ac +++ b/configure.ac @@ -27,6 +27,7 @@ AM_SILENT_RULES([yes]) # Check for programs AC_PROG_CC

[PATCH 4/6] DirectFB integration

2013-10-24 Thread Denis Oliver Kropp
configure: Add DirectFB compositor --- configure.ac | 11 +++ 1 file changed, 11 insertions(+) diff --git a/configure.ac b/configure.ac index 30fb4a3..793da83 100644 --- a/configure.ac +++ b/configure.ac @@ -196,6 +196,17 @@ AS_IF([test x$enable_fbdev_compositor = xyes], [

Re: [RFC 1/5] Add a fullscreen shell protocol

2013-10-24 Thread Jason Ekstrand
On Thu, Oct 24, 2013 at 12:33 PM, Jonas Ådahl jad...@gmail.com wrote: On Wed, Oct 23, 2013 at 06:08:29PM -0500, Jason Ekstrand wrote: Jonas, Thanks for the review! On Wed, Oct 23, 2013 at 3:56 PM, Jonas Ådahl jad...@gmail.com wrote: Hi, Using this protocol, how would the

[RFC v3 1/5] Add a fullscreen shell protocol

2013-10-24 Thread Jason Ekstrand
--- protocol/fullscreen-shell.xml | 61 +++ 1 file changed, 61 insertions(+) create mode 100644 protocol/fullscreen-shell.xml diff --git a/protocol/fullscreen-shell.xml b/protocol/fullscreen-shell.xml new file mode 100644 index 000..9bef555 ---

Re: Buffer release events (was: Add support for eglSwapInterval)

2013-10-24 Thread Jason Ekstrand
This is mostly a recap of the discussion between Neil and myself on IRC this morning. I think a client can guarantee that it gets all of the buffer releases by simply calling wl_display.sync after the attach/damage/commit. It can even delete the wl_callback object immediately without attaching a