Approximately 5% faster on Cortex-A8.
Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit a7878c9f)

Approximately 3% faster on Cortex-A8.
Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit 7da48fd0)

Overlay documentation is still using the old unsupported syntax.
Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit aa8ac53b)

Signed-off-by: Janne Grunau <janne-ffmpeg@jannau.net>
(cherry picked from commit adb1ad0d)

Based on original patch by Kostya Shishkov
Signed-off-by: Janne Grunau <janne-ffmpeg@jannau.net>
(cherry picked from commit e00f41d5)

it's not touched anywhere in ffmpeg, the code setting it was removed
over two years ago (e9b78eeb).
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
(cherry picked from commit b0294c80)

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

While 4 works for some samples, 3 is the correct value since 8x8
DCT is used by (m)jpeg.

Also: cosmetic: split this overly long line

Merge find_near_mvs and mv bitstream decoding: don't do prediction steps
until absolutely necessary.
(cherry picked from commit f3d09d44)

s->windowed_samples will always have a range of [-32767,32767] due to the
window function, so the return value from log2_tab() will always be in the
range [0,14].
Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit 626264b1)

Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit d4582889)

This code will be later split out into a function which takes a 'size'
argument, so I'm keeping the name 'sizeX' here.
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
(cherry picked from commit 0b1d291a)

Only trivial splits are done here -- i.e. copy/paste + reindent +
missing variable declarations.
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
(cherry picked from commit c1fea230)

Signed-off-by: Janne Grunau <janne-ffmpeg@jannau.net>
(cherry picked from commit 68137ba3)

Signed-off-by: Janne Grunau <janne-ffmpeg@jannau.net>
(cherry picked from commit 52091491)

Signed-off-by: Janne Grunau <janne-ffmpeg@jannau.net>
(cherry picked from commit 12c14cd4)

This avoids double semicolons after macro expansion.
Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit 44adbebe)

Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit df211c3a)

(cherry picked from commit 5c19f64c)

Signed-off-by: Janne Grunau <janne-ffmpeg@jannau.net>
(cherry picked from commit 493aa30a)

Instead of returning an error when bytes are left over, just return
the number of actually used bytes as other decoders do.
Instead add a special case so an error will be returned when none
of the data looks valid to avoid making debugging a pain.
Signed-off-by: Janne Grunau <janne-ffmpeg@jannau.net>
(cherry picked from commit 4a72765a)

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
(cherry picked from commit dda3f0ef)

The function return type is void, so a return statement with an
expression is forbidden (and pointless).
Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit b4668274)

Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit dbb09ec2)

Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit 628b16f4)

Its contents aren't used for anything.
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
(cherry picked from commit d7a5106e)

This will be useful for splitting asf_read_header()
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
(cherry picked from commit 7c725380)

This will be useful for splitting asf_read_header()
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
(cherry picked from commit d42b0972)

Keep the original corner case behaviour, where reuse is enabled
for the case where no argument is given to the reuse url option.
Signed-off-by: Martin Storsjö <martin@martin.st>
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
(cherry picked from commit 00952be4)

Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit 569ff021)

The avcodec_thread_free() compatibility wrapper calls ff_thread_free(),
which is not defined when threading is disabled.  Make this call
conditional.
Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit 9a77a92c)

This should fix building with win32 threads.
Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit aef669cd)

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
(cherry picked from commit 4bc328a2)

check AVCodecContext->sample_fmt against AVCodec->sample_fmts[] to ensure
that the encoder supports the specified sample format. Error out if it doesn't.
Previously, it would continue and output garbage. Fixes issue 2587.
(cherry picked from commit 2cfa2d92)

Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit f4c79d1e)

Decode times for big_buck_bunny_720p_stereo:

1 thread:
real    1m14.227s
user    1m13.104s
sys     0m1.108s

2 threads: (33% faster)
real    0m49.329s
user    1m33.735s
sys     0m1.834s

3 threads: (44% faster)
real    0m41.593s
user    1m44.884s
sys     0m1.967s
(cherry picked from commit d23845f3)

(cherry picked from commit f2146944)

As a side effect of the last commit, avcodec_open() now calls it automatically,
so there is no longer any need for clients to call it.
Instead they should set AVCodecContext.thread_count.

avcodec_thread_free() is deprecated, and will be removed from avcodec.h at the
next MAJOR libavcodec bump.

Rename the functions to ff_thread_init/free, since they are now internal.
Wrappers are provided to maintain API compatibility.
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
(cherry picked from commit c0b102ca)

(cherry picked from commit 8e8cc52b)