August 2018 - ffmpeg-cvslog - lists.ffmpeg.org

configure: speed up check_deps()
by Avi Halachmi (:avih) 27 Aug '18

27 Aug '18

ffmpeg | branch: master | Avi Halachmi (:avih) <avihpit(a)yahoo.com> | Sun Aug 5 06:16:44 2018 +0300| [45499e557c808f43175524a98901efeca715813e] | committer: James Almer configure: speed up check_deps() x4 - x25 faster. check_deps() recursively enables/disables components, and its loop is iterated nearly 6000 times. It's particularly slow in bash - currently consuming more than 50% of configure runtime, and about 20% with other shells. This commit applies few local optimizations, most effective first: - Use $1 $2 ... instead of pushvar/popvar, and same at enable_deep* - Abort early in one notable case - empty deps, to avoid costly no-op. - Smaller changes which do add up: - Handle ${cfg}_checking locally instead of via enable[d]/disable - ${cfg}_checking: test done before inprogress - x2 faster in 50%+ - one eval instead of several at the empty-deps early abort path. - The "actual work" part is unmodified - just its surroundings. Biggest speedups (relative and absolute) are observed with bash. Tested-by: Michael Niedermayer <michael(a)niedermayer.cc> Tested-by: Helmut K. C. Tessarek <tessarek(a)evermeet.cx> Tested-by: Dave Yeo <daveryeo(a)telus.net> Tested-by: Reino Wijnsma <rwijnsma(a)xs4all.nl> Signed-off-by: James Almer <jamrial(a)gmail.com> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=45499e557c808f431… --- configure | 86 ++++++++++++++++++++++++++++++++++----------------------------- 1 file changed, 46 insertions(+), 40 deletions(-) diff --git a/configure b/configure index 2f2c2dc58c..ba123eadc0 100755 --- a/configure +++ b/configure @@ -676,13 +676,10 @@ disable_sanitized(){ do_enable_deep(){ for var; do enabled $var && continue - eval sel="\$${var}_select" - eval sgs="\$${var}_suggest" - pushvar var sgs - enable_deep $sel - popvar sgs - enable_deep_weak $sgs - popvar var + set -- $var + eval enable_deep \$${var}_select + var=$1 + eval enable_deep_weak \$${var}_suggest done } @@ -694,9 +691,9 @@ enable_deep(){ enable_deep_weak(){ for var; do disabled $var && continue - pushvar var + set -- $var do_enable_deep $var - popvar var + var=$1 enable_weak $var done } @@ -756,40 +753,49 @@ is_in(){ return 1 } +# The cfg loop is very hot (several thousands iterations), and in bash also +# potentialy quite slow. Try to abort the iterations early, preferably without +# calling functions. 70%+ of the time cfg is already done or without deps. check_deps(){ for cfg; do - enabled ${cfg}_checking && die "Circular dependency for $cfg." - disabled ${cfg}_checking && continue - enable ${cfg}_checking - - eval dep_all="\$${cfg}_deps" - eval dep_any="\$${cfg}_deps_any" - eval dep_con="\$${cfg}_conflict" - eval dep_sel="\$${cfg}_select" - eval dep_sgs="\$${cfg}_suggest" - eval dep_ifa="\$${cfg}_if" - eval dep_ifn="\$${cfg}_if_any" - - pushvar cfg dep_all dep_any dep_con dep_sel dep_sgs dep_ifa dep_ifn - check_deps $dep_all $dep_any $dep_con $dep_sel $dep_sgs $dep_ifa $dep_ifn - popvar cfg dep_all dep_any dep_con dep_sel dep_sgs dep_ifa dep_ifn - - [ -n "$dep_ifa" ] && { enabled_all $dep_ifa && enable_weak $cfg; } - [ -n "$dep_ifn" ] && { enabled_any $dep_ifn && enable_weak $cfg; } - enabled_all $dep_all || { disable $cfg && requested $cfg && die "ERROR: $cfg requested, but not all dependencies are satisfied: $dep_all"; } - enabled_any $dep_any || { disable $cfg && requested $cfg && die "ERROR: $cfg requested, but not any dependency is satisfied: $dep_any"; } - disabled_all $dep_con || { disable $cfg && requested $cfg && die "ERROR: $cfg requested, but some conflicting dependencies are unsatisfied: $dep_con"; } - disabled_any $dep_sel && { disable $cfg && requested $cfg && die "ERROR: $cfg requested, but some selected dependency is unsatisfied: $dep_sel"; } - - enabled $cfg && enable_deep_weak $dep_sel $dep_sgs - - for dep in $dep_all $dep_any $dep_sel $dep_sgs; do - # filter out library deps, these do not belong in extralibs - is_in $dep $LIBRARY_LIST && continue - enabled $dep && eval append ${cfg}_extralibs ${dep}_extralibs - done + eval [ x\$${cfg}_checking = xdone ] && continue + eval [ x\$${cfg}_checking = xinprogress ] && die "Circular dependency for $cfg." + + eval " + dep_all=\$${cfg}_deps + dep_any=\$${cfg}_deps_any + dep_con=\$${cfg}_conflict + dep_sel=\$${cfg}_select + dep_sgs=\$${cfg}_suggest + dep_ifa=\$${cfg}_if + dep_ifn=\$${cfg}_if_any + " + + # most of the time here $cfg has no deps - avoid costly no-op work + if [ "$dep_all$dep_any$dep_con$dep_sel$dep_sgs$dep_ifa$dep_ifn" ]; then + eval ${cfg}_checking=inprogress + + set -- $cfg "$dep_all" "$dep_any" "$dep_con" "$dep_sel" "$dep_sgs" "$dep_ifa" "$dep_ifn" + check_deps $dep_all $dep_any $dep_con $dep_sel $dep_sgs $dep_ifa $dep_ifn + cfg=$1; dep_all=$2; dep_any=$3; dep_con=$4; dep_sel=$5 dep_sgs=$6; dep_ifa=$7; dep_ifn=$8 + + [ -n "$dep_ifa" ] && { enabled_all $dep_ifa && enable_weak $cfg; } + [ -n "$dep_ifn" ] && { enabled_any $dep_ifn && enable_weak $cfg; } + enabled_all $dep_all || { disable $cfg && requested $cfg && die "ERROR: $cfg requested, but not all dependencies are satisfied: $dep_all"; } + enabled_any $dep_any || { disable $cfg && requested $cfg && die "ERROR: $cfg requested, but not any dependency is satisfied: $dep_any"; } + disabled_all $dep_con || { disable $cfg && requested $cfg && die "ERROR: $cfg requested, but some conflicting dependencies are unsatisfied: $dep_con"; } + disabled_any $dep_sel && { disable $cfg && requested $cfg && die "ERROR: $cfg requested, but some selected dependency is unsatisfied: $dep_sel"; } + + enabled $cfg && enable_deep_weak $dep_sel $dep_sgs + + for dep in $dep_all $dep_any $dep_sel $dep_sgs; do + # filter out library deps, these do not belong in extralibs + is_in $dep $LIBRARY_LIST && continue + enabled $dep && eval append ${cfg}_extralibs ${dep}_extralibs + done + fi - disable ${cfg}_checking + eval ${cfg}_checking=done done }

1 0

configure: speed up print_enabled_components()
by Avi Halachmi (:avih) 27 Aug '18

27 Aug '18

ffmpeg | branch: master | Avi Halachmi (:avih) <avihpit(a)yahoo.com> | Mon Jul 30 22:48:04 2018 +0300| [923586a58f37deedeb01f742e4804bc36736a6bc] | committer: James Almer configure: speed up print_enabled_components() x4 - x10 faster. Inside print_enabled components, the filter_list case invokes sed about 350 times to parse the same source file and extract different info for each arg. This is never instant, and on systems where fork is slow (notably MSYS2/Cygwin on windows) it takes many seconds. Change it to use sed once on the source file and set env vars with the parse results, then use these results inside the loop. Additionally, the cases of indev_list and outdev_list are very infrequent, but nevertheless they're faster, and arguably cleaner, with shell parameter substitutions than with command substitutions. Tested-by: Michael Niedermayer <michael(a)niedermayer.cc> Tested-by: Helmut K. C. Tessarek <tessarek(a)evermeet.cx> Tested-by: Dave Yeo <daveryeo(a)telus.net> Tested-by: Reino Wijnsma <rwijnsma(a)xs4all.nl> Signed-off-by: James Almer <jamrial(a)gmail.com> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=923586a58f37deede… --- configure | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/configure b/configure index 95a38cf20e..2f2c2dc58c 100755 --- a/configure +++ b/configure @@ -7237,9 +7237,10 @@ echo "#endif /* AVUTIL_AVCONFIG_H */" >> $TMPH cp_if_changed $TMPH libavutil/avconfig.h -full_filter_name(){ - sed -n "s/^extern AVFilter ff_$[avfsinkrc]\{2,5\}$_$1;/\1_$1/p" $source_path/libavfilter/allfilters.c -} +# full_filter_name_foo=vf_foo +# full_filter_name_bar=asrc_bar +# ... +eval "$(sed -n "s/^extern AVFilter ff_$[avfsinkrc]\{2,5\}$_$.*$;/full_filter_name_\2=\1_\2/p" $source_path/libavfilter/allfilters.c)" # generate the lists of enabled components print_enabled_components(){ @@ -7252,13 +7253,13 @@ print_enabled_components(){ if enabled $c; then case $name in filter_list) - c=$(full_filter_name $(remove_suffix _filter $c)) + eval c=\$full_filter_name_${c%_filter} ;; indev_list) - c=$(add_suffix _demuxer $(remove_suffix _indev $c)) + c=${c%_indev}_demuxer ;; outdev_list) - c=$(add_suffix _muxer $(remove_suffix _outdev $c)) + c=${c%_outdev}_muxer ;; esac printf " &ff_%s,\n" $c >> $TMPH

1 0

configure: speed up flatten_extralibs_wrapper()
by Avi Halachmi (:avih) 27 Aug '18

27 Aug '18

ffmpeg | branch: master | Avi Halachmi (:avih) <avihpit(a)yahoo.com> | Mon Jul 30 22:39:15 2018 +0300| [58b81ac621aea1428aa9e7ca20ad45fc18dba9f7] | committer: James Almer configure: speed up flatten_extralibs_wrapper() x50 - x200 faster. Currently configure spends 50-70% of its runtime inside a single function: flatten_extralibs[_wrapper] - which does string processing. During its run, nearly 20K command substitutions (subshells) are used, including its callees unique() and resolve(), which is the reason for its lengthy run. This commit avoids all subshells during its execution, speeding it up by about two orders of magnitude, and reducing the overall configure runtime by 50-70% . resolve() is rewritten to avoid subshells, and in unique() and flatten_extralibs() we "inline" the filter[_out] functionality. Note that logically, "unique" functionality has more than one possible output (depending on which of the recurring items is kept). As it turns out, other parts expect the last recurring item to be kept (which was the original behavior of uniqie()). This patch preservs its output order. Tested-by: Michael Niedermayer <michael(a)niedermayer.cc> Tested-by: Helmut K. C. Tessarek <tessarek(a)evermeet.cx> Tested-by: Dave Yeo <daveryeo(a)telus.net> Tested-by: Reino Wijnsma <rwijnsma(a)xs4all.nl> Signed-off-by: James Almer <jamrial(a)gmail.com> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=58b81ac621aea1428… --- configure | 46 +++++++++++++++++++++++++++++++++------------- 1 file changed, 33 insertions(+), 13 deletions(-) diff --git a/configure b/configure index b9c9d0b307..95a38cf20e 100755 --- a/configure +++ b/configure @@ -844,22 +844,37 @@ prepend(){ eval "$var=\"$* \$$var\"" } +reverse () { + eval ' + reverse_out= + for v in $'$1'; do + reverse_out="$v $reverse_out" + done + '$1'=$reverse_out + ' +} + +# keeps the last occurence of each non-unique item unique(){ - var=$1 - uniq_list="" - for tok in $(eval echo \$$var); do - uniq_list="$(filter_out $tok $uniq_list) $tok" + unique_out= + eval unique_in=\$$1 + reverse unique_in + for v in $unique_in; do + # " $unique_out" +space such that every item is surrounded with spaces + case " $unique_out" in *" $v "*) continue; esac # already in list + unique_out="$unique_out$v " done - eval "$var=\"${uniq_list}\"" + reverse unique_out + eval $1=\$unique_out } resolve(){ - var=$1 - tmpvar= - for entry in $(eval echo \$$var); do - tmpvar="$tmpvar $(eval echo \$${entry})" + resolve_out= + eval resolve_in=\$$1 + for v in $resolve_in; do + eval 'resolve_out="$resolve_out$'$v' "' done - eval "$var=\"${tmpvar}\"" + eval $1=\$resolve_out } add_cppflags(){ @@ -6734,14 +6749,19 @@ if test $target_os = "haiku"; then fi flatten_extralibs(){ - unset nested_entries + nested_entries= list_name=$1 eval list=\$${1} for entry in $list; do entry_copy=$entry resolve entry_copy - append nested_entries $(filter '*_extralibs' $entry_copy) - flat_entries=$(filter_out '*_extralibs' $entry_copy) + flat_entries= + for e in $entry_copy; do + case $e in + *_extralibs) nested_entries="$nested_entries$e ";; + *) flat_entries="$flat_entries$e ";; + esac + done eval $entry="\$flat_entries" done append $list_name "$nested_entries"

1 0

lavc/encode: fix frame_number double-counted
by Zhong Li 27 Aug '18

27 Aug '18

ffmpeg | branch: master | Zhong Li <zhong.li(a)intel.com> | Thu Jun 28 17:01:46 2018 +0800| [d91370e0f12a3fedd477616011566d9f2fb5e3e5] | committer: Zhong Li lavc/encode: fix frame_number double-counted Encoder frame_number may be double-counted if some frames are cached and then flushed. Take qsv encoder (some frames are cached firsty for asynchronism) as example, ./ffmpeg -loglevel verbose -hwaccel qsv -c:v h264_qsv -i in.mp4 -vframes 100 -c:v h264_qsv out.mp4 frame_number passed to encoder is double-counted and larger than the accurate value. Libx264 encoding with B frames can also reproduce it. Signed-off-by: Zhong Li <zhong.li(a)intel.com> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=d91370e0f12a3fedd… --- libavcodec/encode.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/libavcodec/encode.c b/libavcodec/encode.c index 0ebd8dd6a5..d12c42526b 100644 --- a/libavcodec/encode.c +++ b/libavcodec/encode.c @@ -234,8 +234,8 @@ int attribute_align_arg avcodec_encode_audio2(AVCodecContext *avctx, if (ret >= 0) avpkt->data = avpkt->buf->data; } - - avctx->frame_number++; + if (frame) + avctx->frame_number++; } if (ret < 0 || !*got_packet_ptr) { @@ -329,7 +329,8 @@ int attribute_align_arg avcodec_encode_video2(AVCodecContext *avctx, avpkt->data = avpkt->buf->data; } - avctx->frame_number++; + if (frame) + avctx->frame_number++; } if (ret < 0 || !*got_packet_ptr)

1 0

avcodec/proresdec2: add frame threading support
by Paul B Mahol 26 Aug '18

26 Aug '18

ffmpeg | branch: master | Paul B Mahol <onemda(a)gmail.com> | Wed Aug 22 18:46:51 2018 +0200| [3aacb0d196d1c89c386fe26a5df8d994a3eef619] | committer: Paul B Mahol avcodec/proresdec2: add frame threading support > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=3aacb0d196d1c89c3… --- libavcodec/proresdec2.c | 18 ++++++++++++++++-- 1 file changed, 16 insertions(+), 2 deletions(-) diff --git a/libavcodec/proresdec2.c b/libavcodec/proresdec2.c index b4ea6b5e03..d818e5d8da 100644 --- a/libavcodec/proresdec2.c +++ b/libavcodec/proresdec2.c @@ -36,6 +36,7 @@ #include "simple_idct.h" #include "proresdec.h" #include "proresdata.h" +#include "thread.h" static void permute(uint8_t *dst, const uint8_t *src, const uint8_t permutation[64]) { @@ -644,6 +645,7 @@ static int decode_frame(AVCodecContext *avctx, void *data, int *got_frame, AVPacket *avpkt) { ProresContext *ctx = avctx->priv_data; + ThreadFrame tframe = { .f = data }; AVFrame *frame = data; const uint8_t *buf = avpkt->data; int buf_size = avpkt->size; @@ -669,7 +671,7 @@ static int decode_frame(AVCodecContext *avctx, void *data, int *got_frame, buf += frame_hdr_size; buf_size -= frame_hdr_size; - if ((ret = ff_get_buffer(avctx, frame, 0)) < 0) + if ((ret = ff_thread_get_buffer(avctx, &tframe, 0)) < 0) return ret; decode_picture: @@ -697,6 +699,17 @@ static int decode_frame(AVCodecContext *avctx, void *data, int *got_frame, return avpkt->size; } +#if HAVE_THREADS +static int decode_init_thread_copy(AVCodecContext *avctx) +{ + ProresContext *ctx = avctx->priv_data; + + ctx->slices = NULL; + + return 0; +} +#endif + static av_cold int decode_close(AVCodecContext *avctx) { ProresContext *ctx = avctx->priv_data; @@ -713,7 +726,8 @@ AVCodec ff_prores_decoder = { .id = AV_CODEC_ID_PRORES, .priv_data_size = sizeof(ProresContext), .init = decode_init, + .init_thread_copy = ONLY_IF_THREADS_ENABLED(decode_init_thread_copy), .close = decode_close, .decode = decode_frame, - .capabilities = AV_CODEC_CAP_DR1 | AV_CODEC_CAP_SLICE_THREADS, + .capabilities = AV_CODEC_CAP_DR1 | AV_CODEC_CAP_SLICE_THREADS | AV_CODEC_CAP_FRAME_THREADS, };

1 0

avcodec/mscc: fix several bugs
by Paul B Mahol 26 Aug '18

26 Aug '18

ffmpeg | branch: master | Paul B Mahol <onemda(a)gmail.com> | Sat Apr 22 15:51:52 2017 +0200| [c84a57d1a2189ccd99dc1d5b705f9f1578ff03aa] | committer: Paul B Mahol avcodec/mscc: fix several bugs Fixes #6342. Signed-off-by: Paul B Mahol <onemda(a)gmail.com> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=c84a57d1a2189ccd9… --- libavcodec/mscc.c | 55 +++++++++++++++++++++++++++++++++++++++++-------------- 1 file changed, 41 insertions(+), 14 deletions(-) diff --git a/libavcodec/mscc.c b/libavcodec/mscc.c index 6e4dbb014d..14480cbe13 100644 --- a/libavcodec/mscc.c +++ b/libavcodec/mscc.c @@ -37,10 +37,15 @@ typedef struct MSCCContext { unsigned int uncomp_size; uint8_t *uncomp_buf; z_stream zstream; + + uint32_t pal[256]; } MSCCContext; -static int rle_uncompress(AVCodecContext *avctx, GetByteContext *gb, PutByteContext *pb, int bpp) +static int rle_uncompress(AVCodecContext *avctx, GetByteContext *gb, PutByteContext *pb) { + MSCCContext *s = avctx->priv_data; + unsigned x = 0, y = 0; + while (bytestream2_get_bytes_left(gb) > 0) { uint32_t fill; int j; @@ -78,19 +83,22 @@ static int rle_uncompress(AVCodecContext *avctx, GetByteContext *gb, PutByteCont break; } } + x += run; } else { unsigned copy = bytestream2_get_byte(gb); - if (copy == 1) { + if (copy == 0) { + x = 0; + y++; + bytestream2_seek_p(pb, y * avctx->width * s->bpp, SEEK_SET); + } else if (copy == 1) { return 0; } else if (copy == 2) { - unsigned x, y; - x = bytestream2_get_byte(gb); - y = bytestream2_get_byte(gb); + x += bytestream2_get_byte(gb); + y += bytestream2_get_byte(gb); - bytestream2_skip_p(pb, x * bpp); - bytestream2_skip_p(pb, y * bpp * avctx->width); + bytestream2_seek_p(pb, y * avctx->width * s->bpp + x * s->bpp, SEEK_SET); } else { for (j = 0; j < copy; j++) { switch (avctx->bits_per_coded_sample) { @@ -108,6 +116,10 @@ static int rle_uncompress(AVCodecContext *avctx, GetByteContext *gb, PutByteCont break; } } + + if (s->bpp == 1 && (copy & 1)) + bytestream2_skip(gb, 1); + x += copy; } } } @@ -128,7 +140,8 @@ static int decode_frame(AVCodecContext *avctx, int ret, j; if (avpkt->size < 3) - return AVERROR_INVALIDDATA; + return buf_size; + if ((ret = ff_get_buffer(avctx, frame, 0)) < 0) return ret; @@ -138,6 +151,20 @@ static int decode_frame(AVCodecContext *avctx, buf_size -= 2; } + if (avctx->pix_fmt == AV_PIX_FMT_PAL8) { + int size; + const uint8_t *pal = av_packet_get_side_data(avpkt, AV_PKT_DATA_PALETTE, &size); + + if (pal && size == AVPALETTE_SIZE) { + frame->palette_has_changed = 1; + for (j = 0; j < 256; j++) + s->pal[j] = 0xFF000000 | AV_RL32(pal + j * 4); + } else if (pal) { + av_log(avctx, AV_LOG_ERROR, "Palette size %d is wrong\n", size); + } + memcpy(frame->data[1], s->pal, AVPALETTE_SIZE); + } + ret = inflateReset(&s->zstream); if (ret != Z_OK) { av_log(avctx, AV_LOG_ERROR, "Inflate reset error: %d\n", ret); @@ -156,7 +183,7 @@ static int decode_frame(AVCodecContext *avctx, bytestream2_init(&gb, s->decomp_buf, s->zstream.total_out); bytestream2_init_writer(&pb, s->uncomp_buf, s->uncomp_size); - ret = rle_uncompress(avctx, &gb, &pb, s->bpp); + ret = rle_uncompress(avctx, &gb, &pb); if (ret) return ret; @@ -176,10 +203,10 @@ static int decode_frame(AVCodecContext *avctx, static av_cold int decode_init(AVCodecContext *avctx) { MSCCContext *s = avctx->priv_data; - int zret; + int stride, zret; switch (avctx->bits_per_coded_sample) { - case 8: avctx->pix_fmt = AV_PIX_FMT_GRAY8; break; + case 8: avctx->pix_fmt = AV_PIX_FMT_PAL8; break; case 16: avctx->pix_fmt = AV_PIX_FMT_RGB555; break; case 24: avctx->pix_fmt = AV_PIX_FMT_BGR24; break; case 32: avctx->pix_fmt = AV_PIX_FMT_BGRA; break; @@ -189,13 +216,13 @@ static av_cold int decode_init(AVCodecContext *avctx) } s->bpp = avctx->bits_per_coded_sample >> 3; - memset(&s->zstream, 0, sizeof(z_stream)); + stride = 4 * ((avctx->width * avctx->bits_per_coded_sample + 31) / 32); - s->decomp_size = 4 * avctx->height * ((avctx->width * avctx->bits_per_coded_sample + 31) / 32); + s->decomp_size = 2 * avctx->height * stride; if (!(s->decomp_buf = av_malloc(s->decomp_size))) return AVERROR(ENOMEM); - s->uncomp_size = 4 * avctx->height * ((avctx->width * avctx->bits_per_coded_sample + 31) / 32); + s->uncomp_size = avctx->height * stride; if (!(s->uncomp_buf = av_malloc(s->uncomp_size))) return AVERROR(ENOMEM);

1 0

lavfi/avf_concat: switch to activate.
by Nicolas George 26 Aug '18

26 Aug '18

ffmpeg | branch: master | Nicolas George <george(a)nsup.org> | Fri Aug 10 22:58:46 2018 +0200| [962c9313af31007342dd1ee020a462ef010036cc] | committer: Nicolas George lavfi/avf_concat: switch to activate. Fix trac ticket #7351. > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=962c9313af3100734… --- libavfilter/avf_concat.c | 156 +++++++++++++++++++++++------------------------ 1 file changed, 78 insertions(+), 78 deletions(-) diff --git a/libavfilter/avf_concat.c b/libavfilter/avf_concat.c index 46bd42359b..1d0c2de290 100644 --- a/libavfilter/avf_concat.c +++ b/libavfilter/avf_concat.c @@ -28,8 +28,7 @@ #include "libavutil/channel_layout.h" #include "libavutil/opt.h" #include "avfilter.h" -#define FF_BUFQUEUE_SIZE 256 -#include "bufferqueue.h" +#include "filters.h" #include "internal.h" #include "video.h" #include "audio.h" @@ -48,7 +47,6 @@ typedef struct ConcatContext { int64_t pts; int64_t nb_frames; unsigned eof; - struct FFBufQueue queue; } *in; } ConcatContext; @@ -185,24 +183,6 @@ static int push_frame(AVFilterContext *ctx, unsigned in_no, AVFrame *buf) return ff_filter_frame(outlink, buf); } -static int process_frame(AVFilterLink *inlink, AVFrame *buf) -{ - AVFilterContext *ctx = inlink->dst; - ConcatContext *cat = ctx->priv; - unsigned in_no = FF_INLINK_IDX(inlink); - - if (in_no < cat->cur_idx) { - av_log(ctx, AV_LOG_ERROR, "Frame after EOF on input %s\n", - ctx->input_pads[in_no].name); - av_frame_free(&buf); - } else if (in_no >= cat->cur_idx + ctx->nb_outputs) { - ff_bufqueue_add(ctx, &cat->in[in_no].queue, buf); - } else { - return push_frame(ctx, in_no, buf); - } - return 0; -} - static AVFrame *get_video_buffer(AVFilterLink *inlink, int w, int h) { AVFilterContext *ctx = inlink->dst; @@ -221,11 +201,6 @@ static AVFrame *get_audio_buffer(AVFilterLink *inlink, int nb_samples) return ff_get_audio_buffer(outlink, nb_samples); } -static int filter_frame(AVFilterLink *inlink, AVFrame *buf) -{ - return process_frame(inlink, buf); -} - static void close_input(AVFilterContext *ctx, unsigned in_no) { ConcatContext *cat = ctx->priv; @@ -306,57 +281,10 @@ static int flush_segment(AVFilterContext *ctx) if (ret < 0) return ret; } - /* flush queued buffers */ - /* possible enhancement: flush in PTS order */ - str_max = cat->cur_idx + ctx->nb_outputs; - for (str = cat->cur_idx; str < str_max; str++) { - while (cat->in[str].queue.available) { - ret = push_frame(ctx, str, ff_bufqueue_get(&cat->in[str].queue)); - if (ret < 0) - return ret; - } - } } return 0; } -static int request_frame(AVFilterLink *outlink) -{ - AVFilterContext *ctx = outlink->src; - ConcatContext *cat = ctx->priv; - unsigned out_no = FF_OUTLINK_IDX(outlink); - unsigned in_no = out_no + cat->cur_idx; - unsigned str, str_max; - int ret; - - while (1) { - if (in_no >= ctx->nb_inputs) - return AVERROR_EOF; - if (!cat->in[in_no].eof) { - ret = ff_request_frame(ctx->inputs[in_no]); - if (ret != AVERROR_EOF) - return ret; - close_input(ctx, in_no); - } - /* cycle on all inputs to finish the segment */ - /* possible enhancement: request in PTS order */ - str_max = cat->cur_idx + ctx->nb_outputs - 1; - for (str = cat->cur_idx; cat->nb_in_active; - str = str == str_max ? cat->cur_idx : str + 1) { - if (cat->in[str].eof) - continue; - ret = ff_request_frame(ctx->inputs[str]); - if (ret != AVERROR_EOF) - return ret; - close_input(ctx, str); - } - ret = flush_segment(ctx); - if (ret < 0) - return ret; - in_no += ctx->nb_outputs; - } -} - static av_cold int init(AVFilterContext *ctx) { ConcatContext *cat = ctx->priv; @@ -371,7 +299,6 @@ static av_cold int init(AVFilterContext *ctx) .type = type, .get_video_buffer = get_video_buffer, .get_audio_buffer = get_audio_buffer, - .filter_frame = filter_frame, }; pad.name = av_asprintf("in%d:%c%d", seg, "va"[type], str); if ((ret = ff_insert_inpad(ctx, ctx->nb_inputs, &pad)) < 0) { @@ -387,7 +314,6 @@ static av_cold int init(AVFilterContext *ctx) AVFilterPad pad = { .type = type, .config_props = config_output, - .request_frame = request_frame, }; pad.name = av_asprintf("out:%c%d", "va"[type], str); if ((ret = ff_insert_outpad(ctx, ctx->nb_outputs, &pad)) < 0) { @@ -409,15 +335,88 @@ static av_cold void uninit(AVFilterContext *ctx) ConcatContext *cat = ctx->priv; unsigned i; - for (i = 0; i < ctx->nb_inputs; i++) { + for (i = 0; i < ctx->nb_inputs; i++) av_freep(&ctx->input_pads[i].name); - ff_bufqueue_discard_all(&cat->in[i].queue); - } for (i = 0; i < ctx->nb_outputs; i++) av_freep(&ctx->output_pads[i].name); av_freep(&cat->in); } +static int activate(AVFilterContext *ctx) +{ + ConcatContext *cat = ctx->priv; + AVFrame *frame; + unsigned i, j; + int ret, status; + int64_t pts; + + /* Forward status back */ + for (i = 0; i < ctx->nb_outputs; i++) { + status = ff_outlink_get_status(ctx->outputs[i]); + if (!status) + continue; + for (j = i; j < ctx->nb_inputs; j += ctx->nb_outputs) { + if (!cat->in[j].eof) { + cat->in[j].eof = 1; + ff_inlink_set_status(ctx->inputs[j], status); + return 0; + } + } + + } + + /* Forward available frames */ + if (cat->cur_idx < ctx->nb_inputs) { + for (i = 0; i < ctx->nb_outputs; i++) { + ret = ff_inlink_consume_frame(ctx->inputs[cat->cur_idx + i], &frame); + if (ret < 0) + return ret; + if (ret) { + ff_filter_set_ready(ctx, 10); + return push_frame(ctx, cat->cur_idx + i, frame); + } + } + } + + /* Forward status change */ + if (cat->cur_idx < ctx->nb_inputs) { + for (i = 0; i < ctx->nb_outputs; i++) { + ret = ff_inlink_acknowledge_status(ctx->inputs[cat->cur_idx + i], &status, &pts); + /* TODO use pts */ + if (ret > 0) { + close_input(ctx, cat->cur_idx + i); + if (cat->cur_idx + ctx->nb_outputs >= ctx->nb_inputs) { + ff_outlink_set_status(ctx->outputs[i], status, pts); + } + if (!cat->nb_in_active) { + ret = flush_segment(ctx); + if (ret < 0) + return ret; + } + ff_filter_set_ready(ctx, 10); + return 0; + } + } + } + + ret = FFERROR_NOT_READY; + for (i = 0; i < ctx->nb_outputs; i++) { + if (ff_outlink_frame_wanted(ctx->outputs[i])) { + if (cat->in[cat->cur_idx + i].eof) { + for (j = 0; j < ctx->nb_outputs; j++) + if (!cat->in[cat->cur_idx + j].eof) + ff_inlink_request_frame(ctx->inputs[cat->cur_idx + j]); + return 0; + } else { + ff_inlink_request_frame(ctx->inputs[cat->cur_idx + i]); + ret = 0; + } + } + } + + return ret; +} + static int process_command(AVFilterContext *ctx, const char *cmd, const char *args, char *res, int res_len, int flags) { @@ -437,6 +436,7 @@ AVFilter ff_avf_concat = { .init = init, .uninit = uninit, .query_formats = query_formats, + .activate = activate, .priv_size = sizeof(ConcatContext), .inputs = NULL, .outputs = NULL,

1 0

mpeg4video: Add Studio DPCM support
by Kieran Kunhya 25 Aug '18

25 Aug '18

ffmpeg | branch: master | Kieran Kunhya <kieran(a)kunhya.com> | Sun Aug 19 02:31:42 2018 +0100| [844ff494696e75221e718278139a0db5766ae797] | committer: Paul B Mahol mpeg4video: Add Studio DPCM support > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=844ff494696e75221… --- libavcodec/mpeg4videodec.c | 91 ++++++++++++++++++++++++++++++++++++++++++++-- libavcodec/mpegvideo.c | 73 ++++++++++++++++++++++++++++--------- libavcodec/mpegvideo.h | 2 + 3 files changed, 146 insertions(+), 20 deletions(-) diff --git a/libavcodec/mpeg4videodec.c b/libavcodec/mpeg4videodec.c index 1686ed1001..4c77081237 100644 --- a/libavcodec/mpeg4videodec.c +++ b/libavcodec/mpeg4videodec.c @@ -24,6 +24,7 @@ #include "libavutil/internal.h" #include "libavutil/opt.h" +#include "libavutil/pixdesc.h" #include "error_resilience.h" #include "hwaccel.h" #include "idctdsp.h" @@ -36,6 +37,7 @@ #include "profiles.h" #include "thread.h" #include "xvididct.h" +#include "unary.h" /* The defines below define the number of bits that are read at once for * reading vlc values. Changing these may improve speed and data cache needs @@ -1923,10 +1925,91 @@ static int mpeg4_decode_studio_block(MpegEncContext *s, int32_t block[64], int n return 0; } +static int mpeg4_decode_dpcm_macroblock(MpegEncContext *s, int16_t macroblock[256], int n) +{ + int i, j, w, h, idx = 0; + int block_mean, rice_parameter, rice_prefix_code, rice_suffix_code, + dpcm_residual, left, top, topleft, min_left_top, max_left_top, p, p2, output; + h = 16 >> (n ? s->chroma_y_shift : 0); + w = 16 >> (n ? s->chroma_x_shift : 0); + + block_mean = get_bits(&s->gb, s->avctx->bits_per_raw_sample); + if (block_mean == 0){ + av_log(s->avctx, AV_LOG_ERROR, "Forbidden block_mean\n"); + return AVERROR_INVALIDDATA; + } + s->last_dc[n] = block_mean * (1 << (s->dct_precision + s->intra_dc_precision)); + + rice_parameter = get_bits(&s->gb, 4); + if (rice_parameter == 0) { + av_log(s->avctx, AV_LOG_ERROR, "Forbidden rice_parameter\n"); + return AVERROR_INVALIDDATA; + } + + if (rice_parameter == 15) + rice_parameter = 0; + + if (rice_parameter > 11) { + av_log(s->avctx, AV_LOG_ERROR, "Forbidden rice_parameter\n"); + return AVERROR_INVALIDDATA; + } + + for (i = 0; i < h; i++) { + output = 1 << (s->avctx->bits_per_raw_sample - 1); + top = 1 << (s->avctx->bits_per_raw_sample - 1); + + for (j = 0; j < w; j++) { + left = output; + topleft = top; + + rice_prefix_code = get_unary(&s->gb, 1, 12); + + /* Escape */ + if (rice_prefix_code == 11) + dpcm_residual = get_bits(&s->gb, s->avctx->bits_per_raw_sample); + else { + rice_suffix_code = get_bitsz(&s->gb, rice_parameter); + dpcm_residual = (rice_prefix_code << rice_parameter) + rice_suffix_code; + } + + /* Map to a signed residual */ + if (dpcm_residual & 1) + dpcm_residual = (-1 * dpcm_residual) >> 1; + else + dpcm_residual = (dpcm_residual >> 1); + + if (i != 0) + top = macroblock[idx-w]; + + p = left + top - topleft; + min_left_top = FFMIN(left, top); + if (p < min_left_top) + p = min_left_top; + + max_left_top = FFMAX(left, top); + if (p > max_left_top) + p = max_left_top; + + p2 = (FFMIN(min_left_top, topleft) + FFMAX(max_left_top, topleft)) >> 1; + if (p2 == p) + p2 = block_mean; + + if (p2 > p) + dpcm_residual *= -1; + + macroblock[idx++] = output = (dpcm_residual + p) & ((1 << s->avctx->bits_per_raw_sample) - 1); + } + } + + return 0; +} + static int mpeg4_decode_studio_mb(MpegEncContext *s, int16_t block_[12][64]) { int i; + s->dpcm_direction = 0; + /* StudioMacroblock */ /* Assumes I-VOP */ s->mb_intra = 1; @@ -1945,9 +2028,11 @@ static int mpeg4_decode_studio_mb(MpegEncContext *s, int16_t block_[12][64]) } else { /* DPCM */ check_marker(s->avctx, &s->gb, "DPCM block start"); - avpriv_request_sample(s->avctx, "DPCM encoded block"); - next_start_code_studio(&s->gb); - return SLICE_ERROR; + s->dpcm_direction = get_bits1(&s->gb) ? -1 : 1; + for (i = 0; i < 3; i++) { + if (mpeg4_decode_dpcm_macroblock(s, (*s->dpcm_macroblock)[i], i) < 0) + return AVERROR_INVALIDDATA; + } } if (get_bits_left(&s->gb) >= 24 && show_bits(&s->gb, 23) == 0) { diff --git a/libavcodec/mpegvideo.c b/libavcodec/mpegvideo.c index f75c0fd9b3..d4d3bea649 100644 --- a/libavcodec/mpegvideo.c +++ b/libavcodec/mpegvideo.c @@ -390,6 +390,8 @@ static int init_duplicate_context(MpegEncContext *s) } FF_ALLOCZ_OR_GOTO(s->avctx, s->block32, sizeof(*s->block32), fail) + s->dpcm_direction = 0; + FF_ALLOCZ_OR_GOTO(s->avctx, s->dpcm_macroblock, sizeof(*s->dpcm_macroblock), fail) if (s->avctx->codec_tag == AV_RL32("VCR2")) { // exchange uv @@ -427,6 +429,7 @@ static void free_duplicate_context(MpegEncContext *s) av_freep(&s->me.score_map); av_freep(&s->blocks); av_freep(&s->block32); + av_freep(&s->dpcm_macroblock); av_freep(&s->ac_val_base); s->block = NULL; } @@ -445,6 +448,8 @@ static void backup_duplicate_context(MpegEncContext *bak, MpegEncContext *src) COPY(blocks); COPY(block); COPY(block32); + COPY(dpcm_macroblock); + COPY(dpcm_direction); COPY(start_mb_y); COPY(end_mb_y); COPY(me.map_generation); @@ -820,6 +825,8 @@ static void clear_context(MpegEncContext *s) s->blocks = NULL; s->block32 = NULL; memset(s->pblocks, 0, sizeof(s->pblocks)); + s->dpcm_direction = 0; + s->dpcm_macroblock = NULL; s->ac_val_base = NULL; s->ac_val[0] = s->ac_val[1] = @@ -2132,23 +2139,55 @@ void mpv_reconstruct_mb_internal(MpegEncContext *s, int16_t block[12][64], TODO: Integrate 10-bit properly into mpegvideo.c so that ER works properly */ if (s->avctx->bits_per_raw_sample > 8){ const int act_block_size = block_size * 2; - s->idsp.idct_put(dest_y, dct_linesize, (int16_t*)(*s->block32)[0]); - s->idsp.idct_put(dest_y + act_block_size, dct_linesize, (int16_t*)(*s->block32)[1]); - s->idsp.idct_put(dest_y + dct_offset, dct_linesize, (int16_t*)(*s->block32)[2]); - s->idsp.idct_put(dest_y + dct_offset + act_block_size, dct_linesize, (int16_t*)(*s->block32)[3]); - - dct_linesize = uvlinesize << s->interlaced_dct; - dct_offset = s->interlaced_dct ? uvlinesize : uvlinesize*block_size; - - s->idsp.idct_put(dest_cb, dct_linesize, (int16_t*)(*s->block32)[4]); - s->idsp.idct_put(dest_cr, dct_linesize, (int16_t*)(*s->block32)[5]); - s->idsp.idct_put(dest_cb + dct_offset, dct_linesize, (int16_t*)(*s->block32)[6]); - s->idsp.idct_put(dest_cr + dct_offset, dct_linesize, (int16_t*)(*s->block32)[7]); - if(!s->chroma_x_shift){//Chroma444 - s->idsp.idct_put(dest_cb + act_block_size, dct_linesize, (int16_t*)(*s->block32)[8]); - s->idsp.idct_put(dest_cr + act_block_size, dct_linesize, (int16_t*)(*s->block32)[9]); - s->idsp.idct_put(dest_cb + act_block_size + dct_offset, dct_linesize, (int16_t*)(*s->block32)[10]); - s->idsp.idct_put(dest_cr + act_block_size + dct_offset, dct_linesize, (int16_t*)(*s->block32)[11]); + + if(s->dpcm_direction == 0) { + s->idsp.idct_put(dest_y, dct_linesize, (int16_t*)(*s->block32)[0]); + s->idsp.idct_put(dest_y + act_block_size, dct_linesize, (int16_t*)(*s->block32)[1]); + s->idsp.idct_put(dest_y + dct_offset, dct_linesize, (int16_t*)(*s->block32)[2]); + s->idsp.idct_put(dest_y + dct_offset + act_block_size, dct_linesize, (int16_t*)(*s->block32)[3]); + + dct_linesize = uvlinesize << s->interlaced_dct; + dct_offset = s->interlaced_dct ? uvlinesize : uvlinesize*block_size; + + s->idsp.idct_put(dest_cb, dct_linesize, (int16_t*)(*s->block32)[4]); + s->idsp.idct_put(dest_cr, dct_linesize, (int16_t*)(*s->block32)[5]); + s->idsp.idct_put(dest_cb + dct_offset, dct_linesize, (int16_t*)(*s->block32)[6]); + s->idsp.idct_put(dest_cr + dct_offset, dct_linesize, (int16_t*)(*s->block32)[7]); + if(!s->chroma_x_shift){//Chroma444 + s->idsp.idct_put(dest_cb + act_block_size, dct_linesize, (int16_t*)(*s->block32)[8]); + s->idsp.idct_put(dest_cr + act_block_size, dct_linesize, (int16_t*)(*s->block32)[9]); + s->idsp.idct_put(dest_cb + act_block_size + dct_offset, dct_linesize, (int16_t*)(*s->block32)[10]); + s->idsp.idct_put(dest_cr + act_block_size + dct_offset, dct_linesize, (int16_t*)(*s->block32)[11]); + } + } else if(s->dpcm_direction == 1) { + int i, w, h; + uint16_t *dest_pcm[3] = {(uint16_t*)dest_y, (uint16_t*)dest_cb, (uint16_t*)dest_cr}; + int linesize[3] = {dct_linesize, uvlinesize, uvlinesize}; + for(i = 0; i < 3; i++) { + int idx = 0; + int vsub = i ? s->chroma_y_shift : 0; + int hsub = i ? s->chroma_x_shift : 0; + for(h = 0; h < (16 >> vsub); h++){ + for(w = 0; w < (16 >> hsub); w++) + dest_pcm[i][w] = (*s->dpcm_macroblock)[i][idx++]; + dest_pcm[i] += linesize[i] / 2; + } + } + } else if(s->dpcm_direction == -1) { + int i, w, h; + uint16_t *dest_pcm[3] = {(uint16_t*)dest_y, (uint16_t*)dest_cb, (uint16_t*)dest_cr}; + int linesize[3] = {dct_linesize, uvlinesize, uvlinesize}; + for(i = 0; i < 3; i++) { + int idx = 0; + int vsub = i ? s->chroma_y_shift : 0; + int hsub = i ? s->chroma_x_shift : 0; + dest_pcm[i] += (linesize[i] / 2) * ((16 >> vsub) - 1); + for(h = (16 >> vsub)-1; h >= 1; h--){ + for(w = (16 >> hsub)-1; w >= 1; w--) + dest_pcm[i][w] = (*s->dpcm_macroblock)[i][idx++]; + dest_pcm[i] -= linesize[i] / 2; + } + } } } /* dct only in intra block */ diff --git a/libavcodec/mpegvideo.h b/libavcodec/mpegvideo.h index e16deb64e7..bbc6b5646a 100644 --- a/libavcodec/mpegvideo.h +++ b/libavcodec/mpegvideo.h @@ -509,6 +509,8 @@ typedef struct MpegEncContext { int (*decode_mb)(struct MpegEncContext *s, int16_t block[12][64]); // used by some codecs to avoid a switch() int32_t (*block32)[12][64]; + int dpcm_direction; // 0 = DCT, 1 = DPCM top to bottom scan, -1 = DPCM bottom to top scan + int16_t (*dpcm_macroblock)[3][256]; #define SLICE_OK 0 #define SLICE_ERROR -1

1 0

avformat/mpegts: add missing null pointer checks in ff_parse_mpeg2_descriptor()
by Michael Niedermayer 25 Aug '18

25 Aug '18

ffmpeg | branch: master | Michael Niedermayer <michael(a)niedermayer.cc> | Thu Aug 16 13:33:40 2018 +0200| [323095a6db670041f3c98724c0e9c4a494c55bb9] | committer: Michael Niedermayer avformat/mpegts: add missing null pointer checks in ff_parse_mpeg2_descriptor() Fixes: null pointer dereference Fixes: wtv-crash-75fa58662ded1c1d349f3d1df89394fd690cf92f Found-by: Paul Ch <paulcher(a)icloud.com> Signed-off-by: Michael Niedermayer <michael(a)niedermayer.cc> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=323095a6db670041f… --- libavformat/mpegts.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/libavformat/mpegts.c b/libavformat/mpegts.c index a5cb17ac16..edf6b5701d 100644 --- a/libavformat/mpegts.c +++ b/libavformat/mpegts.c @@ -1983,7 +1983,7 @@ int ff_parse_mpeg2_descriptor(AVFormatContext *fc, AVStream *st, int stream_type int service_type = ((component_type & service_type_mask) >> 3); if (service_type == 0x02 /* 0b010 */) { st->disposition |= AV_DISPOSITION_DESCRIPTIONS; - av_log(ts->stream, AV_LOG_DEBUG, "New track disposition for id %u: %u\n", st->id, st->disposition); + av_log(ts ? ts->stream : fc, AV_LOG_DEBUG, "New track disposition for id %u: %u\n", st->id, st->disposition); } } } @@ -1997,7 +1997,7 @@ int ff_parse_mpeg2_descriptor(AVFormatContext *fc, AVStream *st, int stream_type int service_type = ((component_type & service_type_mask) >> 3); if (service_type == 0x02 /* 0b010 */) { st->disposition |= AV_DISPOSITION_DESCRIPTIONS; - av_log(ts->stream, AV_LOG_DEBUG, "New track disposition for id %u: %u\n", st->id, st->disposition); + av_log(ts ? ts->stream : fc, AV_LOG_DEBUG, "New track disposition for id %u: %u\n", st->id, st->disposition); } } }

1 0

avcodec/hq_hqa: Check remaining input bits in hqa_decode_mb()
by Michael Niedermayer 25 Aug '18

25 Aug '18

ffmpeg | branch: master | Michael Niedermayer <michael(a)niedermayer.cc> | Mon Aug 20 22:53:32 2018 +0200| [c9222b972d6cbdaf6571cf7ae0a6513bffa5ff9f] | committer: Michael Niedermayer avcodec/hq_hqa: Check remaining input bits in hqa_decode_mb() Fixes: Timeout Fixes: 9634/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_HQ_HQA_fuzzer-6267852259590144 Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg Signed-off-by: Michael Niedermayer <michael(a)niedermayer.cc> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=c9222b972d6cbdaf6… --- libavcodec/hq_hqa.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/libavcodec/hq_hqa.c b/libavcodec/hq_hqa.c index f88ad7d5f5..ec9da3e04f 100644 --- a/libavcodec/hq_hqa.c +++ b/libavcodec/hq_hqa.c @@ -181,6 +181,9 @@ static int hqa_decode_mb(HQContext *c, AVFrame *pic, int qgroup, int flag = 0; int i, ret, cbp; + if (get_bits_left(gb) < 1) + return AVERROR_INVALIDDATA; + cbp = get_vlc2(gb, c->hqa_cbp_vlc.table, 5, 1); for (i = 0; i < 12; i++)

1 0