Commit graph

17060 commits

Author SHA1 Message Date
Edward Hervey
63f6f05d66 gst: New Stream listing/selection system
* GstStream
* GstStreamCollection
* GST_EVENT_SELECT_STREAMS
* GST_MESSAGE_STREAM_COLLECTION
2016-06-30 12:31:06 +02:00
Sebastian Dröge
241d0f16f6 poll: #define EWOULDBLOCK to EAGAIN if it's not defined on Windows 2016-06-29 23:24:02 +02:00
Sebastian Dröge
60a2087cb1 bufferpool: Fix handling of the GstPoll
Especially if multiple threads are waiting for buffers to be available again,
the current code was wrong. Fix this and document clearly how the GstPoll is
supposed to be used.

Also fix some potential races with reading from the GstPoll before writing
actually happened.

https://bugzilla.gnome.org/show_bug.cgi?id=767979
2016-06-29 21:21:04 +02:00
Sebastian Dröge
4ff2721810 bus: Make sure to always read the control after popping a message
It might happen that we popped the message before writing of the control
happened. In this case we just have to retry again a bit later, and failure to
do so will cause an additional byte in the control and the GSource /
gst_poll_wait() to always wake up again immediately.

https://bugzilla.gnome.org/show_bug.cgi?id=750397
2016-06-29 21:21:04 +02:00
Sebastian Dröge
149127b755 systemclock: Improve GstPoll handling and don't check for impossible errno values
Also just read/write control every time, GstPoll is optimized by itself
already to only do I/O if switching between empty and one byte.

https://bugzilla.gnome.org/show_bug.cgi?id=750397
2016-06-29 21:21:04 +02:00
Sebastian Dröge
cda3f1213b poll: Clarify when FALSE is returned from read/write_control()
And also mention what the expected values of errno are going to be.

write_control() will only ever return FALSE if there was a critical error. It
will never return because of EINTR, EAGAIN or EWOULDBLOCK.

read_control() will return FALSE if there was no byte to read, in which case
errno would be EWOULDBLOCK.
In all other cases there was a critical error.

https://bugzilla.gnome.org/show_bug.cgi?id=750397
2016-06-29 21:21:04 +02:00
Sebastian Dröge
254955df62 poll: set_controllable(), restart() and set_flushing() are only valid for non-timer GstPolls
On timer GstPolls it will cause the control socket state to become
inconsistent as now one less read_control() than write_control() be would
needed.

Similarly, read_control() and write_control() are only valid on timer
GstPolls.

https://bugzilla.gnome.org/show_bug.cgi?id=750397
2016-06-29 21:21:04 +02:00
Sebastian Dröge
0bfc9fb212 poll: Warn if the return value of gst_poll_read_control() is unused
This might fail even under correct usage, e.g. if read_control() is called
from another thread before write_control() finished in another. It has to be
retried then, or other measures have to be taken, depending on how it is used
by the surrounding code.

https://bugzilla.gnome.org/show_bug.cgi?id=750397
2016-06-29 21:21:04 +02:00
Matthew Gruenke
cd06aea103 poll: Fix various race conditions with read_control() and write_control()
This addresses slightly different race conditions on Linux and Windows, and
fixes gst_poll_read_control() when control_pending == 0.

On Linux, the socketpair() used for control should not be made O_NONBLOCK.
If there's any propagation delay between set->control_write_fd.fd and
set->control_read_fd.fd, even the mutex now held will not be sufficient to
prevent a race condition.  There's no benefit to using O_NONBLOCK, here.
Only liabilities.

For Windows, it's necessary to fix the race condition between testing
set->control_pending and performing WAKE_EVENT()/RELEASE_EVENT().  This is
accomplished by acquiring and holding set->lock, for both of these operations.
We could optimize the Linux version by making this Windows-specific.

For consistency with the Linux implementation, Windows' RELEASE_EVENT()
has also been made to block, although it should never happen.

Also, changed release_wakeup() to return TRUE and decrement control_pending
only when > 0.  Furthermore, RELEASE_EVENT() is called only when
control_pending == 1.

Finally, changed control_pending to use normal, non-atomic arithmetic
operations, since it's now protected by set->lock.

Note: even though the underlying signaling mechanisms are blocking,
release_wakeup() is effectively non-blocking, as it will only attempt to read
from control_read_fd.fd after a byte has been written to control_write_fd.fd
or WaitForSingleObject() after it's been signaled.

https://bugzilla.gnome.org/show_bug.cgi?id=750397
2016-06-29 21:21:04 +02:00
Guillaume Desmottes
e11539c30b bus: chain up GObject::constructed() to the parent class' implementation
Needed so GstBus can be tracked by the leaks tracer.

https://bugzilla.gnome.org/show_bug.cgi?id=768141
2016-06-28 17:23:17 +03:00
Nirbheek Chauhan
54fa564e01 gstconfig.h: Don't use extern with dllexport
GCC emits an error for this with -Werror:

plugin.c:22:1: error: 'gst_plugin_desc' initialized and declared 'extern' [-Werror]

This matches how glib does symbol exporting.

https://bugzilla.gnome.org/show_bug.cgi?id=767463
2016-06-24 01:09:32 +01:00
Nirbheek Chauhan
48088867db win32: Don't use dllexport/import when only building statically
If the prototypes in the public API have dllimport in them when building
statically on Windows, the compiler will look for symbols with symbol
mangling and indirection corresponding to a DLL. This will cause a build
failure when trying to link tests/examples/etc.

External users of GStreamer also need to define -DGST_STATIC_COMPILATION
if they want to link to static gstreamer libraries on Windows.

A similar version of this patch has been committed to all gstreamer
repositories.

https://bugzilla.gnome.org/show_bug.cgi?id=767463
2016-06-23 23:39:45 +01:00
Nicolas Dufresne
084bb505f4 Automatic update of common submodule
From ac2f647 to f363b32
2016-06-21 11:45:26 -04:00
Vincent Penquerc'h
08d30b05c6 tests: add a test for small ring buffer sizes
https://bugzilla.gnome.org/show_bug.cgi?id=767688
2016-06-21 10:20:21 +01:00
Vincent Penquerc'h
b3802f7a9e queue2: fix crash deleting current region for small ring buffers
Ensure we do not attempt to destroy the current range. Doing so
causes the current one to be left dangling, and it may be dereferenced
later, leading to a crash.

This can happen with a very small queue2 ring buffer (10000 bytes)
and 4 kB buffers.

repro case:

gst-launch-1.0 fakesrc sizetype=2 sizemax=4096 ! \
queue2 ring-buffer-max-size=1000 ! fakesink sync=true

https://bugzilla.gnome.org/show_bug.cgi?id=767688
2016-06-21 10:20:20 +01:00
Tim-Philipp Müller
e452acb634 tests: gstobject: fix typo in test name 2016-06-20 11:34:49 +01:00
Reynaldo H. Verdejo Pinochet
6b8494072e docs/design/part-tracing: fix reference to renamed func 2016-06-17 13:31:12 -07:00
Nicolas Dufresne
4603722e2d tee: Properly handle return value when only 1 pad
This patch handle the case when you have 1 pad (so the fast path is
being used) but this pad is removed. If we are in allow-not-linked, we
should return GST_FLOW_OK, otherwise, we should return GST_FLOW_UNLINKED
and ignore the meaningless return value obtained from pushing.

https://bugzilla.gnome.org/show_bug.cgi?id=767413
2016-06-17 13:03:20 +03:00
Stefan Sauer
06d2c6989d gst-plot-traces.sh: add a script to plot gst-tracer graphs
The script extracts cpu-usage data from a tracelog and plots it via gnuplot.
2016-06-16 15:53:52 +02:00
Sebastian Dröge
6b9e70d193 device: Fix typo
paramater -> parameter
2016-06-15 16:12:36 +02:00
Tim-Philipp Müller
09d5ff4097 info: flesh out GST_PTR_FORMAT docs a bit 2016-06-14 19:16:33 +01:00
Sebastian Dröge
b9a4a2a952 basesink: Update start time when losing state only if we were in PLAYING
If we were in PAUSED, the current clock time and base time don't have much to
do with the running time anymore as the clock might have advanced while we
were PAUSED. The system clock does that for example, audio clocks often don't.

Updating the start time in PAUSED will cause a) the wrong position to be
reported, b) step events to step not just the requested amount but the amount
of time we spent in PAUSED. The start time should only ever be updated when
going from PLAYING to PAUSED to remember the current running time (to be able
to compensate later when going to PLAYING for the clock time advancing while
PAUSED), not when we are already in PAUSED.

Based on a patch by Kishore Arepalli <kishore.arepalli@gmail.com>

The updating of the start time when the state is lost was added in commit
ba943a82c0 to fix the position reporting when
the state is lost. This still works correctly after this change.

https://bugzilla.gnome.org/show_bug.cgi?id=739289
2016-06-13 20:20:44 +02:00
Sebastian Dröge
37713c3388 pad: Log pad offsets as signed times 2016-06-11 22:18:22 +03:00
Sebastian Dröge
da46fe236b pad: Also check the number of segment events and if other serialized events and queries trigger segment updating too
https://bugzilla.gnome.org/show_bug.cgi?id=765049
2016-06-11 21:57:00 +03:00
Sebastian Dröge
c8cae6a94b pad: Add unit test for pad offset handling on src pads
https://bugzilla.gnome.org/show_bug.cgi?id=765049
2016-06-11 21:40:15 +03:00
Sebastian Dröge
8c7da1d426 adapter: Rename functions and implement new functions, update test
We don't do calculations with different units (buffer offsets and bytes)
anymore but have functions for:
1) getting the number of bytes since the last discont
2) getting the offset (and pts/dts) at the last discont

and the previously added function to get the last offset and its distance from
the current adapter position.

https://bugzilla.gnome.org/show_bug.cgi?id=766647
2016-06-10 09:49:33 +03:00
Edward Hervey
67ae0ad225 adapter: Add methods to query current offset
API: gst_buffer_prev_offset
API: gst_buffer_get_offset_from_discont

The gst_buffer_get_offset_from_discont() method allows retrieving the current
offset based on the GST_BUFFER_OFFSET of the buffers that were pushed in.

The offset will be set initially by the GST_BUFFER_OFFSET of
DISCONT buffers, and then incremented by the sizes of the following
buffers.

The gst_buffer_prev_offset() method allows retrievent the previous
GST_BUFFER_OFFSET regardless of flags. It works in the same way as
the other gst_buffer_prev_*() methods.

https://bugzilla.gnome.org/show_bug.cgi?id=766647
2016-06-10 09:49:33 +03:00
Tim-Philipp Müller
665de91347 gstconfig.h.in: indent #if #else jungle for better readability 2016-06-09 17:48:40 +01:00
Sebastian Dröge
1dbb27f3a7 utils: Add gst_pad_link_maybe_ghosting() for consistency
We already had a _full() version, but having that alone seems inconsistent.
Add a non-full version that mirrors the behaviour of gst_pad_link() vs
gst_pad_link_full().
2016-06-08 12:12:28 +03:00
Edward Hervey
ea395c2498 baseparse: Make sure DISCONT flags are properly propagated
If we drop a frame that contained a discontinuity, we must remember
that for the next frame that *will* be pushed downstream.

https://bugzilla.gnome.org/show_bug.cgi?id=766795
2016-06-07 09:42:39 +02:00
Tim-Philipp Müller
56b9290073 deviceprovider: remove base_class_finalize function
It's not going to get called anyway.

https://bugzilla.gnome.org/show_bug.cgi?id=765540
2016-06-04 13:35:12 +01:00
Tim-Philipp Müller
38c74e41c7 element: remove base_class_finalize_func which is never called
Won't be called for static types, so no point keeping it around.

https://bugzilla.gnome.org/show_bug.cgi?id=765540
2016-06-04 13:11:55 +01:00
Tim-Philipp Müller
6446a1a45c tracers: leaks: some micro-optimisations
- we know number of filter items is not going to change,
  but compiler doesn't

- only do GST_IS_TRACER check for GObjects, not mini objects

- use non-type check cast macros in performance critical paths
2016-06-03 14:03:25 +01:00
Guillaume Desmottes
0bb7fa9855 tracers: add leaks tracer
https://bugzilla.gnome.org/show_bug.cgi?id=765052
2016-06-03 00:36:46 +01:00
Guillaume Desmottes
4a41468ce7 Use MAY_BE_LEAKED_FLAG
This helps having "make check" passing with the leaks tracer enabled.

https://bugzilla.gnome.org/show_bug.cgi?id=766008
2016-06-02 23:14:15 +01:00
Guillaume Desmottes
21070c8114 tracing: add hooks when objects or miniobjects are created and destroyed
https://bugzilla.gnome.org/show_bug.cgi?id=765052
2016-06-02 23:10:44 +01:00
Guillaume Desmottes
194964b4e0 gst_deinit: move down tracers cleaning
We want the tracer detecting leaks to be finalized as late as possible
to give the chance to other gst components to be properly cleaned first.

https://bugzilla.gnome.org/show_bug.cgi?id=765052
2016-06-02 23:08:04 +01:00
Guillaume Desmottes
d479eefde0 tests: plugin: remove feature refcount assert
This check fails if one, or more, tracers are loaded while running the
test. The new "leaks" tracer will be able to check for leaks anyway.

https://bugzilla.gnome.org/show_bug.cgi?id=765052
2016-06-02 23:05:34 +01:00
Guillaume Desmottes
aa336008b3 tracerrecord: allow G_TYPE_POINTER for field types
Tracers may want to display the address of an object.

https://bugzilla.gnome.org/show_bug.cgi?id=765052
2016-06-02 22:53:28 +01:00
Stefan Sauer
f2fd3bda2b gstobject: split up name tests
It is better to have separate tests:
1) the test name will tell what is broekn when the test fails
2) we still run the other tests when one assert fails
3) the tests are easier to understand
4) we don't rely on sie effect of previous actions
5) ...

Also ix the assertion message for the name checks (Gst -> fakeobject).
2016-05-30 13:45:02 +02:00
Stefan Sauer
1385e24bc6 design: update design doc
Some of the api was renamed before the merge.
2016-05-30 02:06:01 -07:00
Stefan Sauer
b23cb42ae9 docs: xref the free function and expand allocation query docs
Add xrefs for how to parse pool details from an allocation query.
2016-05-30 02:04:18 -07:00
Nicolas Dufresne
850510f9e8 object: Add _set_name() test on parented object
This is not allowed, and set_name() should fail.

https://bugzilla.gnome.org/show_bug.cgi?id=766923
2016-05-26 15:36:52 -04:00
Nicolas Dufresne
1a5f79981c object: Check that name change are notified once
GObject allow calling g_object_notify() within set_property() and
won't notify it twice. As it was raised during review, add a unit test to
make sure.

https://bugzilla.gnome.org/show_bug.cgi?id=766923
2016-05-26 15:36:21 -04:00
Nicolas Dufresne
446778464b object: Notify name change when using _set_name()
There was a 0.11 FIXME about notifying the name change or removing that
function. Clearly we can't remove this function, so let's notify it.

https://bugzilla.gnome.org/show_bug.cgi?id=766923
2016-05-26 15:36:05 -04:00
Edward Hervey
6bfb88b410 gst_private: Fix gstconfig include
Since it's a generated header, we need to specify the gst subdir so
that it gets properly included in out-of-dir compilation
2016-05-25 15:31:52 +02:00
Tim-Philipp Müller
5628f7ea73 gst: make sure to include gstconfig.h also in gst_private.h
For GST_EXPORT define and also things like GST_DISABLE_REGISTRY.
Hopefully fixes the following build failure on cerbero-cross-mingw32:
helpers/gst-plugin-scanner.c:50: undefined reference to `_imp___gst_disable_registry_cache'
2016-05-25 10:48:05 +01:00
Tim-Philipp Müller
0274650e42 g-i: pass compiler env to g-ir-scanner
It's what introspection.mak does as well. Should
fix spurious build failures on gnome-continuous.
2016-05-24 00:40:27 +01:00
Tim-Philipp Müller
af47500d12 gst: g-i: pass compiler with quotes
So CC="ccache gcc" works properly.
2016-05-23 21:15:48 +01:00
Ray Strode
4152f266f7 gst: attempt to fix/track-down mysterious gnome-continuous build failures 2016-05-23 21:08:07 +01:00