Signed-off-by: Anton Khirnov <anton@khirnov.net>

This restores the performance of simple frame-threaded codecs back to
their pre-merge levels.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Coccinelle profile used:

  @@
  expression r, ctx, f, loglevel, str, flags;
  @@

  -if ((r = ff_get_buffer(ctx, f, flags)) < 0) {
  -    av_log(ctx, loglevel, str);
  -    return r;
  -}
  +if ((r = ff_get_buffer(ctx, f, flags)) < 0)
  +    return r;

  @@
  expression r, ctx, f, loglevel, str;
  @@

  -if ((r = ff_reget_buffer(ctx, f)) < 0) {
  -    av_log(ctx, loglevel, str);
  -    return r;
  -}
  +if ((r = ff_reget_buffer(ctx, f)) < 0)
  +    return r;

  @@
  expression r, ctx, f, loglevel, str, flags;
  @@

  -if ((r = ff_thread_get_buffer(ctx, f, flags)) < 0) {
  -    av_log(ctx, loglevel, str);
  -    return r;
  -}
  +if ((r = ff_thread_get_buffer(ctx, f, flags)) < 0)
  +    return r;

...along with some manual patches for the remaining ones.

Fixes memleaks when frame mt is used and the decoder is not flushed at
the end.

I.e. don't do anything on already released frames.

Fixes assertion failure

Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

This fixes the following warning with GCC:
    libavcodec/pthread.c:35:0: warning: _GNU_SOURCE redefined [enabled by default]
    <command-line>::0: note: this is the location of the previous definition

The reason of the presence of this flag is:
    % pkg-config --cflags sdl
    -D_GNU_SOURCE=1 -D_REENTRANT -I/usr/include/SDL

It will be useful in the upcoming transition to refcounted AVFrames.

This might fix a hypothetical memleak
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

pthread_wait_cond can wake up for no reason (Wikipedia: Spurious_wakeup).
The FF_THREAD_SLICE thread mechanism could spontaneously execute jobs or
allow the caller of avctx->execute to return before all jobs were complete.
This adds tests to both cases to ensure the wakeup is real.
Signed-off-by: Ben Jackson <ben@ben.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Mans Rullgard <mans@mansr.com>

Signed-off-by: Martin Storsjö <martin@martin.st>

No speed difference was meassureable
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

We generally do not declare variables within for statements and
there are compilers that choke on such constructs.

Signed-off-by: Diego Biurrun <diego@biurrun.de>

This removes all references to AVCodecContext.dsp_mask and marks
it for eviction at the next version bump.  It has been superseded
by av_set_cpu_flag_mask() which, unlike this field, works everywhere.
Signed-off-by: Mans Rullgard <mans@mansr.com>

corrects invalid return when avpkt->size == 0.

Reproducible with test case and ffplay -threads 2. Stack trace:
http://pastebin.com/PexZ4Uc0

Test case:
http://commondatastorage.googleapis.com/dalecurtis-shared/crash.ogmSigned-off-by: Dale Curtis <dalecurtis@chromium.org>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Reproducible with test case and ffplay -threads 2. Stack trace:
http://pastebin.com/PexZ4Uc0

Test case:
http://commondatastorage.googleapis.com/dalecurtis-shared/crash.ogmSigned-off-by: Dale Curtis <dalecurtis@chromium.org>
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

Reviewed-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Before this, they were only added to the delayed release queue and not
freed until later. This could lead to unnecessary memory use or buffer
exhaustion.
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

Else we run out of progress variables after a few failed buffer
allocations.
Signed-off-by: Anton Khirnov <anton@khirnov.net>

Signed-off-by: Anton Khirnov <anton@khirnov.net>

This fixes race conditions that ultimately lead to memory corruption.

Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Calling avcodec_flush_buffers() and then avcodec_decode_video2() with
a 0-sized packet (to get remaining buffered frames) could incorrectly
return an old frame from before the avcodec_flush_buffers() call. Add
a loop in ff_thread_flush() to zero the got_frame field of each thread
to ensure the old frames will not be returned.
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

In most places where it's used, it's as a pointless write-only field.

Only rv10 decoder actually reads from it, but it stores some internal
version info in it. There is no reason for it to be in a public field.

Put related fields together, move the important ones closer to the
beginning, try to eliminate holes.

The cleanup is only done now when
a picture is returned (assuming that it has to be done when its returned)
a error is returned (assuming that there will be no further progress on the frame)
the codec is not h264 (this is still needed due to some deadlocks in realvideo)

This fixes a decoding regression with 00017.MTS
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

This order is more logic and might prevent a race.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Fixes bug118, bug120 and bug125 at least
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

This fixes memory corruption when seeking in broken streams.
a random mpeg4 in nut file was used to debug.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

have been seen waiting and deadlocking on it in bug125.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

They are used to signal the number of slices and offsets of each slice
out of band to the decoder.