Before this patch, we explicitly modify rsp, which isn't necessarily
universally acceptable, since the space under the stack pointer might
be modified in things like signal handlers. Therefore, use an explicit
register to hold the stack pointer relative to the bottom of the stack
(i.e. rsp). This will also clear out valgrind errors about the use of
uninitialized data that started occurring after the idct16x16/ssse3
optimizations were first merged.

Cycle measurements for intra itxfm_4x4_add on ped1080p.webm:
idct_idct:    66 -> 67 cycles (noise measurement)
idct_iadst:  199 -> 79 cycles
iadst_idct:  165 -> 70 cycles
iadst_iadst: 183 -> 82 cycles

Cycle measurements for intra itxfm_8x8_add on ped1080p.webm:
idct_idct:   133 -> 135 cycles (noise measurement)
idct_iadst:  900 -> 241 cycles
iadst_idct:  864 -> 215 cycles
iadst_iadst: 973 -> 310 cycles

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

The muxer will write at least the number of bytes requested and possibly
up to 3 bytes more.  This is because the muxer writes 32-bit integers
and the format requires 4-byte alignment anyway.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

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

As with the change to flacenc this allows the user to control the amount
of padding they want added to the file.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

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

Allows a user to control the amount, if any, of padding they want added
to the file.  If set to zero the block will not be written at all.  If
set to some positive number four more bytes will be added to the file
due to the small header required for the block.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

This field is used to store the number of bytes that should be written
as padding to a metadata header of a file.  For example:
 - The FLAC format's METADATA_BLOCK_PADDING [1]
 - The ID3v2 tag format's padding           [2]

[1] http://xiph.org/flac/format.html#metadata_block_padding
[2] http://id3.org/id3v2.3.0#ID3v2_overviewSigned-off-by: Michael Niedermayer <michaelni@gmx.at>

* Using extradata by TS demuxer to store values from PMT
* Support of multiple languages in one DVB teletext stream:
  comma separated language codes in metadata "language" field
Reviewed-by: Marton Balint <cus@passwd.hu>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

* Using extradata by TS muxer to correctly restore PMT table
* PES_header_data_length should be always 0x24 for DVB teletext,
  according to DVB standard
Reviewed-by: Marton Balint <cus@passwd.hu>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Reviewed-by: Marton Balint <cus@passwd.hu>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: João Bernardo Oliveira <jbvsmo@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

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

Mpeg1/2 should not need it
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

This is needed in case the checked bitstream reader is disabled
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Prevents some overreads at the cost of 1 cpu cycle
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

sandybridge i7 274->260 cycles
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

No speedloss meassured
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

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

* qatar/master:
  vp9: fix bugs in updating coef probabilities with parallelmode=1

Conflicts:
	libavcodec/vp9.c

No change, the bug/buggy code doesnt seem to be in FFmpeg
Merged-by: Michael Niedermayer <michaelni@gmx.at>

* commit 'bd014123':
  vp9: fix mvref finding to adhere to bug in libvpx.

Conflicts:
	libavcodec/vp9mvs.c

No change
See: dd3d0f46Merged-by: Michael Niedermayer <michaelni@gmx.at>

It is implied by av_register_all(). Simplify.

Avoid doc duplication and desynching.

- The memcpy was completely wrong because
s->prob_ctx[s->framectxid].coef is a [4][2][2][6][6][3] array, whereas
s->prob.coef is a [4][2][2][6][6][11] array.
- The additional check was committed to ffmpeg by Ronald S. Bultje.
Signed-off-by: Anton Khirnov <anton@khirnov.net>

Fixes a particular youtube video that I unfortunately can't share.
Signed-off-by: Anton Khirnov <anton@khirnov.net>

Fixes CID1163848
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Fixes CID1163852
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Fixes CID1163850
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

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

This avoids the need for double error checking
Fixes CID1163843
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Fixes CID1163842
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Found-by: Dale Curtis <dalecurtis@google.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Reynaldo H. Verdejo Pinochet <r.verdejo@sisa.samsung.com>

All samples in the wild are RGB565/555 and we are already
acting on this assumption when pushing out the video frames,
so if we get anything != than 16 for bpp we just override
this value for doing any calculations making our approach
consistent.

Also avoid repeatedly shifting bpp.
Signed-off-by: Reynaldo H. Verdejo Pinochet <r.verdejo@sisa.samsung.com>

MTV has a 512 bytes header. We should be able to
parse required fiels from the first 57 (rest is
padding) but if we haven't seen enough data then
go with SCORE_EXTENSION.
Signed-off-by: Reynaldo H. Verdejo Pinochet <r.verdejo@sisa.samsung.com>

Abort immediately if we are not getting enough
data to extract the required fields.
Signed-off-by: Reynaldo H. Verdejo Pinochet <r.verdejo@sisa.samsung.com>

Now it is possible to adjust compression speed vs R/D when needed
and also skip vintage players compatibility at will.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>