If after computing the suggestion with downstream caps we still have
a non-fixed suggestion caps try to intersect with the input caps
of the pad alloc to avoid useless renegotiations.
https://bugzilla.gnome.org/show_bug.cgi?id=642130
This can happen for example when downstream proposed new caps, later proposed
the previous caps again which in turn enables passthrough mode in upstream
elements and the wrong-sized buffer appears in an element where the caps
change never happened. Simply allocate a new buffer in this case.
See bug #635461.
Make the _get_caps functions behave like the _get_caps_reffed variants and
remove the _reffed variants. This means that _get_caps doesn't return a writable
caps anymore and an explicit _make_writable() is needed before modifying the
caps.
Because of the awkward refcounting in prepare_output_buffer, we might end up
with writable buffers that point to the same data. Check for those cases so that
we avoid a useless memcpy and keep valgrind quiet.
Fixes#628176
When we are handling a buffer and need to allocate an output buffer, handle the
case when downstream suggests us a format that we can't convert the input buffer
to. In that case, check if there is another format available downstream instead
of failing.
Fixes#621332 and see also #614296
If initially pass-through caps are negotiated between a transform element's
sink and src pads, but then the downstream element returns different caps
on a buffer from pad_alloc(), basetransform gets stuck with proxy_alloc=TRUE
even though the upstream peer doesn't accept the caps, causing
gst_pad_peer_accept_caps() to be called on each buffer in _buffer_alloc():
if (!gst_caps_is_equal (newcaps, caps)) {
GST_DEBUG_OBJECT (trans, "caps are new");
/* we have new caps, see if we can proxy downstream */
>> if (gst_pad_peer_accept_caps (pad, newcaps)) {
/* peer accepts the caps, return a buffer in this format */
GST_DEBUG_OBJECT (trans, "peer accepted new caps");
which is taking ~40ms/frame.
This patch does two things. (1) if the buffer returned from pad_alloc() has
new caps, trigger the decision whether to proxy the buffer-alloc to be
revisited, and (2) disable proxy if peer does not accept new caps. (The first
part may not be strictly needed, but seemed like a good idea.)
Note that this issue would not arise except in case of downstream elements
who have on their template-caps, some that would be suitable for pass-through,
but at runtime pick more restrictive caps (for ex, after querying a driver for
what formats it actually supports).
When basetransform received an unsupported caps on pad_alloc
it just returned not-negotiated. This patch makes it query
the allowed caps between his sinkpad and upstream's srcpad
to find a caps to suggest.
This happens when dinamically switching pipeline elements
and upstream pad_allocs with the previous caps that was
being used.
Fixes#614296
Allow subclasses to override the acceptcaps function because in some cases a
custom implementation can be much much faster than the default one.
See #621190
When doing pad_allocs, use non-fixed caps suggestions and
try to fixate them before using. This makes possible to
have suggested buffer size with 0 in basetransform just
to signal upstream a renegotiation is needed
Fixes#576234Fixes#609046
There's not much point in using GST_DEBUG_FUNCPTR with GObject
virtual functions such as get_property, set_propery, finalize and
dispose, since they'll never be used by anyone anyway. Saves a
few bytes and possibly a tenth of a polar bear.
In most places in core and baseclasses we just need the caps to do caps-
intersections. In that case ref'ed caps are enough (no need to copy).
This patch also switches the code to use the new functions.
API: gst_pad_get_caps_refed(), gst_pad_peer_get_caps_refed()
Check when we need to touch the metadata of the output buffer after selecting
the output buffer so that we have everything in one place.
Also take flags and timestamp modifications into account.
When we have the same input as output caps, reuse the input caps object. After
the caps refcounting has been sorted out now, we can finally enable this
optimisation.
While reconfiguring a basetransform element we need also to recheck
the alloc request. Because it's possible that due to caps changes
the proxy_alloc state is not correct anymore.
(Re-commit after discusion with Wim on IRC)
While reconfiguring a basetransform element we need also to recheck
the alloc request. Because it's possible that due to caps changes
the proxy_alloc state is not correct anymore.
Original commit message from CVS:
* gst/gstdebugutils.c:
Add some ideas, how to make the graph smaller.
* gst/gstutils.c:
Add a comment from a debug session.
* libs/gst/base/gstbasetransform.c:
Log more context.
* libs/gst/controller/gstinterpolationcontrolsource.c:
Indet.
* plugins/elements/gstcapsfilter.c:
Fix typo in docs.
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_prepare_output_buffer),
(gst_base_transform_getrange), (gst_base_transform_chain):
* libs/gst/base/gstbasetransform.h:
Add vmethod that is called before we start the transform and which can
be used to configure the transform, such as dynamic properties.
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_acceptcaps):
Revert quick accepcaps attempt, it's not fully equivalent to the old
behaviour and thus causes regressions.
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c: (gst_base_transform_init),
(gst_base_transform_getcaps), (gst_base_transform_find_transform),
(gst_base_transform_acceptcaps), (gst_base_transform_getrange):
Add beginnings of a more optimized acceptcaps function than the default
core one.
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_find_transform), (gst_base_transform_getrange):
If we have a fixate function, call it even if we already have fixed caps
because the subclass might add some caps. Makes audioconvert add a
default channel layout.
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_prepare_output_buffer),
(gst_base_transform_getrange):
Clear the output buffer variable.
Cleanups to the error path in the getrange function.
Fixes#557649.
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_prepare_output_buffer),
(gst_base_transform_buffer_alloc), (gst_base_transform_suggest):
Protect sink_alloc caps with the sinkpad lock to avoid nasty caps
refcount problems as seen in banshee and maybe also in farsight2.
Remove atomic int now that we need to take the lock anyways.
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c:
* plugins/elements/gstcapsfilter.c:
* tests/check/Makefile.am:
* tests/check/elements/.cvsignore:
* tests/check/elements/capsfilter.c:
Fix assertion in basetransform when the subclass chooses not to
allocate a buffer in prepare_buffer(), and make capsfilter error out
cleanly if requested to apply caps that don't completely specify the
buffer. Fixes#551509
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_prepare_output_buffer):
Take new caps ref because our old one might have been gone when the
subclass performs a gst_pad_set_caps() on the srcpad. See #548764.
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_prepare_output_buffer):
If the element is configured in passthrough mode but the
prepare_output_buffer gave us a new output buffer, discard that buffer
and reuse the input buffer.
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_prepare_output_buffer),
(gst_base_transform_buffer_alloc):
Go over the buffer_alloc function again and make sure we always end up
allocating a buffer.
Add some more docs.
Avoid doing pad alloc when we have a pending suggestion because we
cannot yet deal with changing caps in that case. Fixes#547728
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_transform_caps),
(gst_base_transform_prepare_output_buffer),
(gst_base_transform_buffer_alloc), (gst_base_transform_suggest):
Don't overwrite the outsize when calculating the expected size of a new
buffer because we still need it in case we cannot process the new
buffer.
When converting the size of the new buffer to an upstream size, actually
use the expected size of the buffer, not some other random value.
Use an atomic int to signal that a new upstream caps suggestion is
available.
When we can convert the current buffer to a new format, check if the
buffer size is of the expected size and allocate a new buffer of the
expected size when this is not the case.
* tests/check/libs/transform1.c: (GST_START_TEST):
remove ifdeffed code from the unit test.
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_class_init), (gst_base_transform_init),
(gst_base_transform_transform_caps),
(gst_base_transform_prepare_output_buffer):
Add debug info.
When a buffer is writable, its metadata is also writable so we don't
need to subbuffer (which then makes the buffer not-writable anymore).
Original commit message from CVS:
Patch by: Tim-Philipp Müller <tim.muller at collabora co uk>
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_sink_event):
Unref events that the GstBaseTransform::event vfunc didn't want to
have forwarded by the base class. Closes a leak in identity.
Fixes bug #446763.
Original commit message from CVS:
Patch by: Sjoerd Simons <sjoerd at luon dot net>
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_buffer_alloc):
Check the caps of the buffer returned by gst_pad_alloc_buffer() and
fall back to default negotiation in the chain function if the caps
are different from what was requested. Fixes bug #526768.
Original commit message from CVS:
Based on a patch by: Sjoerd Simons <sjoerd at luon dot net>
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_buffer_alloc):
Don't passthrough buffer allocation too easily if the caps change.
This breaks when working in passthrough mode and upstream changes
it's caps. Fixes bug #526768.
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_prepare_output_buffer):
Also unset the GAP flag on buffers if we're working inplace but
the element is not GAP-aware.
Mark a comment as FIXME 0.11.
Original commit message from CVS:
Patch by: Mark Nauwelaerts <manauw at skynet be>
* gst/gstclock.h:
* libs/gst/base/gstbasesrc.h:
* libs/gst/base/gstbasetransform.c:
* libs/gst/check/gstcheck.c:
Small documentation fixes. Fixes#523978.
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_transform_size):
Print element name with g_warning() if there's a problem
with the unit size.
Original commit message from CVS:
* docs/libs/gstreamer-libs-sections.txt:
* libs/gst/base/gstbasetransform.c: (gst_base_transform_init),
(gst_base_transform_prepare_output_buffer),
(gst_base_transform_set_gap_aware):
* libs/gst/base/gstbasetransform.h:
API: Add gst_base_transform_set_gap_aware() to control whether
the element correctly handles GST_BUFFER_FLAG_GAP or shouldn't
get buffers with this flag at all. Fixes#503231.
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_sink_event):
Work around the brokenness of the event vmethod in basetransform. Prefer
to return TRUE when the subclass returned FALSE (meaning don't forward
the event).
* libs/gst/base/gstbasetransform.h:
Clarify the docs.
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c: (gst_base_transform_chain):
Properly set the last_stop position on GstSegment. This will only happen
if there is a buffer to push out.
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_buffer_alloc):
always_in_place does not mean that the sink and source caps are the
same! Make sure we don't blindly proxy the buffer_alloc in this case.
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c: (gst_base_transform_activate):
Make sure streaming has finished before calling the ::stop() vfunc,
since that vfunc might clear state which is being used in the
streaming thread. This fixes a race that caused crashes in
audioresample when shutting down a pipeline (#420106).
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_sink_eventfunc),
(gst_base_transform_handle_buffer), (gst_base_transform_chain),
(gst_base_transform_activate):
* libs/gst/base/gstbasetransform.h:
Add support for dropping buffers with custom GstFlowReturn.
Set DISCONT flags on outgoing buffers based on QoS, incomming DISCONT
buffers or dropped buffers.
* docs/libs/gstreamer-libs-sections.txt:
docs for new custom return code.
* plugins/elements/gstidentity.c: (gst_identity_transform_ip):
Use drop support in base class to implement drop-probability.
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c: (gst_base_transform_init),
(gst_base_transform_check_get_range):
Implement a checkgetrange function instead of relying on the default
core behaviour that assumes we can operate in pull mode if we have a
getrange function. First step at fixing #385084.
Original commit message from CVS:
2007-01-12 Andy Wingo <wingo@pobox.com>
* libs/gst/base/gstbasetransform.c (gst_base_transform_activate):
Factor start/stop into this private function instead of partially
in activate functions and partially in the change_state function.
Fixes setup before the element has changed from READY->PAUSED, as
is the case in pull-mode pipelines.
(gst_base_transform_sink_activate_push)
(gst_base_transform_src_activate_pull): Refactor to use
gst_base_transform_activate().
(gst_base_transform_change_state): Removed, not needed any more.
* libs/gst/base/gstbasesink.c (gst_base_sink_negotiate_pull):
Truncate before fixating.
Original commit message from CVS:
2007-01-12 Andy Wingo <wingo@pobox.com>
* libs/gst/base/gstbasetransform.c (_GstBaseTransformPrivate):
(gst_base_transform_init, gst_base_transform_sink_activate_push)
(gst_base_transform_src_activate_pull):
Track the activation mode.
(gst_base_transform_setcaps): In pull mode, when activating the
src pad, after activating the sink pad, activate the sink pad's
peer, as discussed in part-negotiation.txt.
* libs/gst/base/gstbasesrc.h:
* libs/gst/base/gstbasesrc.c (gst_base_src_fixate): Add fixate
vmethod, as in basesink.
* libs/gst/base/gstbasesink.h: Reformat docs, add fixate vmethod.
* libs/gst/base/gstbasesink.c (gst_base_sink_pad_setcaps): In pull
mode, first proxy the setcaps to the peer pad.
(gst_base_sink_pad_fixate): Add a fixate function that calls the
new fixate vmethod.
(gst_base_sink_default_activate_pull): Rename from
gst_base_sink_activate_pull.
(gst_base_sink_negotiate_pull): New function, performs negotiation
in pull mode before calling ::activate_pull().
(gst_base_sink_pad_activate_pull): Actually call the activate_pull
vmethod instead of the default implementation. I have no idea how
this worked before. Negotiate before calling activate_pull.
Original commit message from CVS:
* gst/gstevent.h:
Add small comment.
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_sink_eventfunc):
Debug segment values *after* updating them as this is more
interesting.
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_buffer_alloc):
Check if requested caps are the same as the sinks caps IF
->have_same_caps is TRUE. If they are not, act as if have_same_caps
is FALSE.
This fixes the renegotiation issues stated in #352827.
Original commit message from CVS:
* docs/gst/gstreamer-sections.txt:
* gst/gstcaps.c: (gst_structure_is_equal_foreach),
(gst_caps_merge):
* gst/gstcaps.h:
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_transform_caps):
API: Add gst_caps_merge() and use it in basetransform, fixes#345444
in a better way
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_buffer_alloc):
Only call downstream buffer_alloc if transform element is passthrough
or always_in_place. Closes#350449.
Original commit message from CVS:
* gst/gst.c:
* gst/gstpad.c: (gst_pad_set_active):
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_transform_caps):
Simplify caps to get rid of duplicates, fixes#345444
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_buffer_alloc):
Even if we can't figure out the proper format to request downstream,
call buffer_alloc() downstream with the input parameters without setting
the caps on the srcpad. This will force negotiation in the chain
function.
Closes#350449
Original commit message from CVS:
* libs/gst/base/gstbasetransform.c:
(gst_base_transform_buffer_alloc):
Use OBJECT_LOCK and refcounting to get the pad caps in the
buffer_alloc function because the caps could change while we are
busy with them. Fixes#349105
