Signed-off-by: Timothy Gu <timothygu99@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

also remove spinlock, it doesnt work on AIX

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

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

Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>

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

UPDATE_CACHE isnt needed and can cause segfaults

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

Should fix Ticket2825

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

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

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

Signed-off-by: Paul B Mahol <onemda@gmail.com>

similar is already done in the other get_bits() functions

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

This allows more transparent mixing of get_bits and whole-byte access
without having to touch get_bits internals.

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

This allows more transparent mixing of get_bits and whole-byte access
without having to touch get_bits internals.

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

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

Too much code relies in having init_get_bits fed with a valid
buffer and set its dimension to 0.

Check for NULL buffer instead.

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

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

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

Accept the buffer size in bytes and check for overflow before passing
the value in bits to init_get_bits.

Fix an undefined behaviour and make the function return a proper
error in case of overflow.

CC: libav-stable@libav.org

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

Also remove a duplicate function in the MPEG-TS demuxer.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Justin Ruggles <justin.ruggles@gmail.com>

also the 0 case is indeed used in some code pathes

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

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

This fixes a theoretical signed overflow

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

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

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

The prefix is a historic remnant that probably meant "alternative".
Now that the A32 bitstream reader has been dropped it makes no sense anymore.

that way qatar maintains the code for me and i dont need to resolve conflicts.
If someone wants the a32 reader back, only thing you need to do is maintain
it, i would be happy to have it back, iam just not volunteering to maintain
it due to lack of time.

Based on: a1e98f198e by Mans Rullgard.

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

This macro is empty since the removal of the A32 bitstream reader.

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

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

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

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

When turned on, H264/CAVLC gets ~15% (CVPCMNL1_SVA_C.264) slower for
ultra-high-bitrate files, or ~2.5% (CVFI1_SVA_C.264) for lower-bitrate
files. Other codecs are affected to a lesser extent because they are
less optimized; e.g., VC-1 slows down by less than 1% (all on x86).
The patch generated 3 extra instructions (cmp, cmovae and mov) per
call to get_bits().

The performance penalty on ARM is within the error margin for most
files, up to 4% in extreme cases such as CVPCMNL1_SVA_C.264.

Based on work (for GCI) by Aneesh Dogra <lionaneesh@gmail.com>, and
inspired by patch in Chromium by Chris Evans <cevans@chromium.org>.

The A32 bitstream reader variant is only used on ARMv5 and for
Prores due to the larger bit cache this decoder requires.

In benchmarks on ARMv5 (Marvell Sheeva) with gcc 4.6, the only
statistically significant difference between ALT and A32 is
a 4% advantage for ALT in FLAC decoding.  There is thus no (longer)
any reason to keep the A32 reader from this point of view.

This patch adds an option to the ALT reader increasing the bit
cache to 32 bits as required by the Prores decoder.  Benchmarking
shows no significant change in speed on Intel i7.  Again, the
A32 reader fails to justify its existence.

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

These pointers are already of the correct type.

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

This reverts commit 23ce6e7212.

x86 does not use this variant so having inline asm there
is pointless.

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

This silences numerous compiler warnings from skip_bits(),
where the cache variable is not used.

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

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

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

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

Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit 611a6f59ce)

These whitespace changes improve the readability of the get_bits
macros.

Signed-off-by: Mans Rullgard <mans@mansr.com>
(cherry picked from commit fb5c841d5f)