A lot of changes happen at the same time:
- Add a framequeue fifo to AVFilterLink.
- split AVFilterLink.status into status_in and status_out: requires
changes to the few filters and programs that use it directly
(f_interleave, split, filtfmts).
- Add a field ready to AVFilterContext, marking when the filter is ready
and its activation priority.
- Add flags to mark blocked links.
- Change ff_filter_frame() to enqueue the frame.
- Change all filtering functions to update the ready field and the
blocked flags.
- Update ff_filter_graph_run_once() to use the ready field.
- buffersrc: always push the frame immediately.
It makes more sense to not claim anything about the SAR if we don't know
anything. No changes in the FATE tests, since this is what avconv ends
up doing anyway.
Many of the functions from avfilter/formats can return errors, usually AVERROR(ENOMEM).
This propagates the return values.
All of these were found by using av_warn_unused_result, demonstrating its utility.
Tested with FATE. I am least sure of the changes to avfilter/filtergraph,
since I don't know what/how reduce_format is intended to behave and how it should
react to errors.
Fixes: CID 1325680, 1325679, 1325678.
Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Previous version Reviewed-by: Nicolas George <george@nsup.org>
Previous version Reviewed-by: Clément Bœsch <u@pkh.me>
Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
Previously it was partly only added to central functions called
internally, however if GCC would inline these into the public fuction, the
alignment attribute would not take effect.
Instead, add it to all public entry points to avoid these problems.
Signed-off-by: Martin Storsjö <martin@martin.st>
Previously it was partly only added to central functions called
internally, however if GCC would inline these into the public fuction, the
alignment attribute would not take effect.
Instead, add it to all public entry points to avoid these problems.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
The functions which actually drive the filter graph by pushing
frames through it need to ensure an aligned stack for SIMD functions.
This fixes a crash in YADIF filter when using a mingw build in a MSVC
application.
Signed-off-by: Martin Storsjö <martin@martin.st>
The line was lost during the AVFrame transition.
It causes programs relying on the failed requests count,
especially ffmpeg, to add frames to the wrong input.
Fix trac ticket #2467.
Number of planes is not always equal to the number of components even
for formats marked with PIX_FMT_PLANAR -- e.g. NV12 has three components
in two planes.
The functions which actually drive the filter graph by pushing
frames through it need to ensure an aligned stack for SIMD functions.
This fixes a crash in YADIF filter when using a mingw build in a MSVC application.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
The PUSH flags is necessary for efficient scheduling;
otherwise there is no feedback when adding frames to
closed paths.
The NO_CHECK_FORMAT is a small optimization that does
not cost much to implement.
The KEEP_REF flag maps to the add/write distinction in
the fork's API.
If buffersrc was configured for frames with an unknown layout,
the incoming frames will have channel_layout = 0.
If the format negotiation has selected a known (and compatible)
channel layout for the link, the frame is assumed to have
that layout, the field must be set before injecting the frame
in the filters.
Nicolas George
Anton Khirnov
Nicolas George
Ganesh Ajjanagadde
Vittorio Giovara
Hendrik Leppkes
Federico Tomassetti
Michael Niedermayer
Federico Tomassetti
Clément Bœsch
Lukasz Marek
Stefano Sabatini
Diego Biurrun
Paul B Mahol
Clément Bœsch