No buffer will be fetched from the pool after it's uninitialized, so there's
no benefit from waiting until every single buffer has been returned to it
before freeing them all.
This should free some memory in certain scenarios, which can be beneficial in
low memory systems.
Based on a patch by Jonas Karlman.
Reviewed-by: Anton Khirnov <anton@khirnov.net>
Signed-off-by: James Almer <jamrial@gmail.com>
The last user of g15Mask, r15Mask, g16Mask and r16Mask was disabled
in 77a416e8aa and finally removed in
36e8de07ed62609df45d064b56501e3084d25723; b15Mask and b16Mask were
apparently always unused (except for in_asm_used_var_warning_killer,
a function that only existed to make the compiler not optimize ASM
constants away).
w10 is unused since d604bab901f6dfaaad672ef2164e42b1f350474c, w02
since ef423a6618.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
mask24hh etc. are unused since f099fbf5f3ac1d6b3753fc8dfda6558572111fbd,
mask32b and mask32r since 296609f859a587575b91fe9e9691f2707d6e8136,
mask32g since b38d487466 and mask32 since
f8a138be52.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
mmx_grnmask is unused since 531f97b0c32d1d421f3ac614e002c53951658115,
the other constants since e934194b6a.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
The code using ff_exp2 (namely ff_acelp_decode_gain_code) use it only if
G729_BITEXACT is defined. So disable it if not.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
This allowed to remove forward declarations. Because compilers expect
declarations for all functions they encounter even when it is within
blocks disabled via "if (0 && foo)", one has to use a real #if in
ff_diracdsp_init_x86.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Unused since a1f3b18bf55f106c974eacb1dc831be4d2bd5277, yet as nonstatic
functions the compiler can't detect this, so that these functions aren't
stripped and no warning is emitted.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
ff_acelp_decode_8bit_to_1st_delay3, ff_acelp_decode_4bit_to_2nd_delay3
and ff_acelp_decode_5_6_bit_to_2nd_delay3 are all only used once (by
g729dec) whereas ff_acelp_decode_9bit_to_1st_delay6 and
ff_acelp_decode_6bit_to_2nd_delay6 are completely unused; with the
possible exception of ff_acelp_decode_4bit_to_2nd_delay3, these
functions are so small that inlining them is appropriate; and as long as
ff_acelp_decode_4bit_to_2nd_delay3 is only called once, this is also
true for it.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Added in 0c79b1402a to use it
in a function that was never used and was itself removed in
676f1f533e.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Clang infers from the existence of a default case that said case can be
taken. In case of libavcodec/bitstream.c said default case consisted of
an av_assert1 that evaluates to nothing in case of the ordinary assert
level. In this case (that doesn't happen) a variable wouldn't be
initialized, so Clang emitted Wsometimes-uninitialized warnings.
Solve this by making sure that the default path also initializes
the aforementioned variable.
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
The counter for the number of styles is written on two bytes, ergo
anything > UINT16_MAX is invalid. This also fixes a compiler warning
because of a tautologically true check on 64bit systems.
Reviewed-by: Philip Langdale <philipl@overt.org>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Commit 0d1229f1d2 factored the main part
of the voc demuxer's read_packet function out; yet when this Libav
commit was merged in f99195d56f4aab266926724ca1cfae822df4df16, the
dependency of the other users of this function on vocdec.o was
unnecessarily kept. This commit fixes this.
While just at it, also disable the data only used by the voc demuxer
and muxer in voc.c if both of them are disabled.
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
tiff.c is the only user of the data from tiff_data.c (the dependency of
the tiff encoder of it is spurious). Therefore this commit moves all the
data from tiff_data.c to tiff_data.h (which is only included by tiff.c)
and makes the objects declared therein static.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
The options of the w64 demuxer are a proper subset of the options for
the wav demuxer, making it possible to reuse a part of the options for
the wav demuxer for the w64 demuxer.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
The NUT and avi demuxers only need ff_codec_movvideo_tags and so this
removes a dependency on the rest of isom.c as well as on mpeg4audio.c
(which isom depends on); it is similar for the Matroska demuxer and
muxers, except that the mpeg4audio.c dependency can't be avoided.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
This is a result of the mov channel parsing stuff being factored out
of mov.c twice: Once in 91b782720f
to isom.c and later in 3bab7cd128.
Also remove the isom.h header; and while just at it, remove an unused
mathematics.h inclusion.
(isom.c actually depends upon mpeg4audio from libavcodec for
avpriv_mpeg4audio_get_config2 and avpriv_mpa_freq_tab; yet there is
no configure dependency for iso_media which leads to failure of shared
builds.)
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
It only existed because some code in mjpegenc_common.c relied on it;
yet said code was actually only used by mjpegenc.c and has been moved
there.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@gmail.com>
Anton Khirnov
Paul B Mahol
James Almer
Andreas Rheinhardt