1
0
mirror of https://github.com/mpv-player/mpv synced 2024-12-26 00:42:57 +00:00
mpv/common/av_common.c
Philip Langdale 4574dd5dc6 ffmpeg: update to handle deprecation of av_init_packet
This has been a long standing annoyance - ffmpeg is removing
sizeof(AVPacket) from the API which means you cannot stack-allocate
AVPacket anymore. However, that is something we take advantage of
because we use short-lived AVPackets to bridge from native mpv packets
in our main decoding paths.

We don't think that switching these to `av_packet_alloc` is desirable,
given the cost of heap allocation, so this change takes a different
approach - allocating a single packet in the relevant context and
reusing it over and over.

That's fairly straight-forward, with the main caveat being that
re-initialising the packet is unintuitive. There is no function that
does exactly what we need (what `av_init_packet` did). The closest is
`av_packet_unref`, which additionally frees buffers and side-data.
However, we don't copy those things - we just assign them in from our
own packet, so we have to explicitly clear the pointers before calling
`av_packet_unref`. But at least we can make a wrapper function for
that.

The weirdest part of the change is the handling of the vtt subtitle
conversion. This requires two packets, so I had to pre-allocate two in
the context struct. That sounds excessive, but if allocating the
primary packet is too expensive, then allocating the secondary one for
vtt subtitles must also be too expensive.

This change is not conditional as heap allocated AVPackets were
available for years and years before the deprecation.
2022-12-03 14:44:18 -08:00

419 lines
13 KiB
C

/*
* This file is part of mpv.
*
* mpv is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* mpv is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with mpv. If not, see <http://www.gnu.org/licenses/>.
*/
#include <assert.h>
#include <math.h>
#include <limits.h>
#include <libavutil/common.h>
#include <libavutil/log.h>
#include <libavutil/dict.h>
#include <libavutil/opt.h>
#include <libavutil/error.h>
#include <libavutil/cpu.h>
#include <libavcodec/avcodec.h>
#include <libavformat/avformat.h>
#include "config.h"
#include "audio/chmap_avchannel.h"
#include "common/common.h"
#include "common/msg.h"
#include "demux/packet.h"
#include "demux/stheader.h"
#include "misc/bstr.h"
#include "video/fmt-conversion.h"
#include "av_common.h"
#include "codecs.h"
int mp_lavc_set_extradata(AVCodecContext *avctx, void *ptr, int size)
{
if (size) {
av_free(avctx->extradata);
avctx->extradata_size = 0;
avctx->extradata = av_mallocz(size + AV_INPUT_BUFFER_PADDING_SIZE);
if (!avctx->extradata)
return -1;
avctx->extradata_size = size;
memcpy(avctx->extradata, ptr, size);
}
return 0;
}
enum AVMediaType mp_to_av_stream_type(int type)
{
switch (type) {
case STREAM_VIDEO: return AVMEDIA_TYPE_VIDEO;
case STREAM_AUDIO: return AVMEDIA_TYPE_AUDIO;
case STREAM_SUB: return AVMEDIA_TYPE_SUBTITLE;
default: return AVMEDIA_TYPE_UNKNOWN;
}
}
AVCodecParameters *mp_codec_params_to_av(struct mp_codec_params *c)
{
AVCodecParameters *avp = avcodec_parameters_alloc();
if (!avp)
return NULL;
// If we have lavf demuxer params, they overwrite by definition any others.
if (c->lav_codecpar) {
if (avcodec_parameters_copy(avp, c->lav_codecpar) < 0)
goto error;
return avp;
}
avp->codec_type = mp_to_av_stream_type(c->type);
avp->codec_id = mp_codec_to_av_codec_id(c->codec);
avp->codec_tag = c->codec_tag;
if (c->extradata_size) {
uint8_t *extradata = c->extradata;
int size = c->extradata_size;
if (avp->codec_id == AV_CODEC_ID_FLAC) {
// ffmpeg expects FLAC extradata to be just the STREAMINFO,
// so grab only that (and assume it'll be the first block)
if (size >= 8 && !memcmp(c->extradata, "fLaC", 4)) {
extradata += 8;
size = MPMIN(34, size - 8); // FLAC_STREAMINFO_SIZE
}
}
avp->extradata = av_mallocz(size + AV_INPUT_BUFFER_PADDING_SIZE);
if (!avp->extradata)
goto error;
avp->extradata_size = size;
memcpy(avp->extradata, extradata, size);
}
avp->bits_per_coded_sample = c->bits_per_coded_sample;
// Video only
avp->width = c->disp_w;
avp->height = c->disp_h;
// Audio only
avp->sample_rate = c->samplerate;
avp->bit_rate = c->bitrate;
avp->block_align = c->block_align;
#if !HAVE_AV_CHANNEL_LAYOUT
avp->channels = c->channels.num;
if (!mp_chmap_is_unknown(&c->channels))
avp->channel_layout = mp_chmap_to_lavc(&c->channels);
#else
mp_chmap_to_av_layout(&avp->ch_layout, &c->channels);
#endif
return avp;
error:
avcodec_parameters_free(&avp);
return NULL;
}
// Set avctx codec headers for decoding. Returns <0 on failure.
int mp_set_avctx_codec_headers(AVCodecContext *avctx, struct mp_codec_params *c)
{
enum AVMediaType codec_type = avctx->codec_type;
enum AVCodecID codec_id = avctx->codec_id;
AVCodecParameters *avp = mp_codec_params_to_av(c);
if (!avp)
return -1;
int r = avcodec_parameters_to_context(avctx, avp) < 0 ? -1 : 0;
avcodec_parameters_free(&avp);
if (avctx->codec_type != AVMEDIA_TYPE_UNKNOWN)
avctx->codec_type = codec_type;
if (avctx->codec_id != AV_CODEC_ID_NONE)
avctx->codec_id = codec_id;
return r;
}
// Pick a "good" timebase, which will be used to convert double timestamps
// back to fractions for passing them through libavcodec.
AVRational mp_get_codec_timebase(struct mp_codec_params *c)
{
AVRational tb = {c->native_tb_num, c->native_tb_den};
if (tb.num < 1 || tb.den < 1) {
if (c->reliable_fps)
tb = av_inv_q(av_d2q(c->fps, 1000000));
if (tb.num < 1 || tb.den < 1)
tb = AV_TIME_BASE_Q;
}
// If the timebase is too coarse, raise its precision, or small adjustments
// to timestamps done between decoder and demuxer could be lost.
if (av_q2d(tb) > 0.001) {
AVRational r = av_div_q(tb, (AVRational){1, 1000});
tb.den *= (r.num + r.den - 1) / r.den;
}
av_reduce(&tb.num, &tb.den, tb.num, tb.den, INT_MAX);
if (tb.num < 1 || tb.den < 1)
tb = AV_TIME_BASE_Q;
return tb;
}
static AVRational get_def_tb(AVRational *tb)
{
return tb && tb->num > 0 && tb->den > 0 ? *tb : AV_TIME_BASE_Q;
}
// Convert the mpv style timestamp (seconds as double) to a libavcodec style
// timestamp (integer units in a given timebase).
int64_t mp_pts_to_av(double mp_pts, AVRational *tb)
{
AVRational b = get_def_tb(tb);
return mp_pts == MP_NOPTS_VALUE ? AV_NOPTS_VALUE : llrint(mp_pts / av_q2d(b));
}
// Inverse of mp_pts_to_av(). (The timebases must be exactly the same.)
double mp_pts_from_av(int64_t av_pts, AVRational *tb)
{
AVRational b = get_def_tb(tb);
return av_pts == AV_NOPTS_VALUE ? MP_NOPTS_VALUE : av_pts * av_q2d(b);
}
// Set dst from mpkt. Note that dst is not refcountable.
// mpkt can be NULL to generate empty packets (used to flush delayed data).
// Sets pts/dts using mp_pts_to_av(ts, tb). (Be aware of the implications.)
// Set duration field only if tb is set.
void mp_set_av_packet(AVPacket *dst, struct demux_packet *mpkt, AVRational *tb)
{
dst->side_data = NULL;
dst->side_data_elems = 0;
dst->buf = NULL;
av_packet_unref(dst);
dst->data = mpkt ? mpkt->buffer : NULL;
dst->size = mpkt ? mpkt->len : 0;
/* Some codecs (ZeroCodec, some cases of PNG) may want keyframe info
* from demuxer. */
if (mpkt && mpkt->keyframe)
dst->flags |= AV_PKT_FLAG_KEY;
if (mpkt && mpkt->avpacket) {
dst->side_data = mpkt->avpacket->side_data;
dst->side_data_elems = mpkt->avpacket->side_data_elems;
if (dst->data == mpkt->avpacket->data)
dst->buf = mpkt->avpacket->buf;
dst->flags |= mpkt->avpacket->flags;
}
if (mpkt && tb && tb->num > 0 && tb->den > 0)
dst->duration = mpkt->duration / av_q2d(*tb);
dst->pts = mp_pts_to_av(mpkt ? mpkt->pts : MP_NOPTS_VALUE, tb);
dst->dts = mp_pts_to_av(mpkt ? mpkt->dts : MP_NOPTS_VALUE, tb);
}
void mp_set_avcodec_threads(struct mp_log *l, AVCodecContext *avctx, int threads)
{
if (threads == 0) {
threads = av_cpu_count();
if (threads < 1) {
mp_warn(l, "Could not determine thread count to use, defaulting to 1.\n");
threads = 1;
} else {
mp_verbose(l, "Detected %d logical cores.\n", threads);
if (threads > 1)
threads += 1; // extra thread for better load balancing
}
// Apparently some libavcodec versions have or had trouble with more
// than 16 threads, and/or print a warning when using > 16.
threads = MPMIN(threads, 16);
}
mp_verbose(l, "Requesting %d threads for decoding.\n", threads);
avctx->thread_count = threads;
}
static void add_codecs(struct mp_decoder_list *list, enum AVMediaType type,
bool decoders)
{
void *iter = NULL;
for (;;) {
const AVCodec *cur = av_codec_iterate(&iter);
if (!cur)
break;
if (av_codec_is_decoder(cur) == decoders &&
(type == AVMEDIA_TYPE_UNKNOWN || cur->type == type))
{
mp_add_decoder(list, mp_codec_from_av_codec_id(cur->id),
cur->name, cur->long_name);
}
}
}
void mp_add_lavc_decoders(struct mp_decoder_list *list, enum AVMediaType type)
{
add_codecs(list, type, true);
}
// (Abuses the decoder list data structures.)
void mp_add_lavc_encoders(struct mp_decoder_list *list)
{
add_codecs(list, AVMEDIA_TYPE_UNKNOWN, false);
}
char **mp_get_lavf_demuxers(void)
{
char **list = NULL;
void *iter = NULL;
int num = 0;
for (;;) {
const AVInputFormat *cur = av_demuxer_iterate(&iter);
if (!cur)
break;
MP_TARRAY_APPEND(NULL, list, num, talloc_strdup(NULL, cur->name));
}
MP_TARRAY_APPEND(NULL, list, num, NULL);
return list;
}
int mp_codec_to_av_codec_id(const char *codec)
{
int id = AV_CODEC_ID_NONE;
if (codec) {
const AVCodecDescriptor *desc = avcodec_descriptor_get_by_name(codec);
if (desc)
id = desc->id;
if (id == AV_CODEC_ID_NONE) {
const AVCodec *avcodec = avcodec_find_decoder_by_name(codec);
if (avcodec)
id = avcodec->id;
}
}
return id;
}
const char *mp_codec_from_av_codec_id(int codec_id)
{
const char *name = NULL;
const AVCodecDescriptor *desc = avcodec_descriptor_get(codec_id);
if (desc)
name = desc->name;
if (!name) {
const AVCodec *avcodec = avcodec_find_decoder(codec_id);
if (avcodec)
name = avcodec->name;
}
return name;
}
bool mp_codec_is_lossless(const char *codec)
{
const AVCodecDescriptor *desc =
avcodec_descriptor_get(mp_codec_to_av_codec_id(codec));
return desc && (desc->props & AV_CODEC_PROP_LOSSLESS);
}
// kv is in the format as by OPT_KEYVALUELIST(): kv[0]=key0, kv[1]=val0, ...
// Copy them to the dict.
void mp_set_avdict(AVDictionary **dict, char **kv)
{
for (int n = 0; kv && kv[n * 2]; n++)
av_dict_set(dict, kv[n * 2 + 0], kv[n * 2 + 1], 0);
}
// For use with libav* APIs that take AVDictionaries of options.
// Print options remaining in the dict as unset.
void mp_avdict_print_unset(struct mp_log *log, int msgl, AVDictionary *dict)
{
AVDictionaryEntry *t = NULL;
while ((t = av_dict_get(dict, "", t, AV_DICT_IGNORE_SUFFIX)))
mp_msg(log, msgl, "Could not set AVOption %s='%s'\n", t->key, t->value);
}
// If the name starts with "@", try to interpret it as a number, and set *name
// to the name of the n-th parameter.
static void resolve_positional_arg(void *avobj, char **name)
{
if (!*name || (*name)[0] != '@' || !avobj)
return;
char *end = NULL;
int pos = strtol(*name + 1, &end, 10);
if (!end || *end)
return;
const AVOption *opt = NULL;
int offset = -1;
while (1) {
opt = av_opt_next(avobj, opt);
if (!opt)
return;
// This is what libavfilter's parser does to skip aliases.
if (opt->offset != offset && opt->type != AV_OPT_TYPE_CONST)
pos--;
if (pos < 0) {
*name = (char *)opt->name;
return;
}
offset = opt->offset;
}
}
// kv is in the format as by OPT_KEYVALUELIST(): kv[0]=key0, kv[1]=val0, ...
// Set these options on given avobj (using av_opt_set..., meaning avobj must
// point to a struct that has AVClass as first member).
// Options which fail to set (error or not found) are printed to log.
// Returns: >=0 success, <0 failed to set an option
int mp_set_avopts(struct mp_log *log, void *avobj, char **kv)
{
return mp_set_avopts_pos(log, avobj, avobj, kv);
}
// Like mp_set_avopts(), but the posargs argument is used to resolve positional
// arguments. If posargs==NULL, positional args are disabled.
int mp_set_avopts_pos(struct mp_log *log, void *avobj, void *posargs, char **kv)
{
int success = 0;
for (int n = 0; kv && kv[n * 2]; n++) {
char *k = kv[n * 2 + 0];
char *v = kv[n * 2 + 1];
resolve_positional_arg(posargs, &k);
int r = av_opt_set(avobj, k, v, AV_OPT_SEARCH_CHILDREN);
if (r == AVERROR_OPTION_NOT_FOUND) {
mp_err(log, "AVOption '%s' not found.\n", k);
success = -1;
} else if (r < 0) {
char errstr[80];
av_strerror(r, errstr, sizeof(errstr));
mp_err(log, "Could not set AVOption %s='%s' (%s)\n", k, v, errstr);
success = -1;
}
}
return success;
}
/**
* Must be used to free an AVPacket that was used with mp_set_av_packet().
*
* We have a particular pattern where we "borrow" buffers and set them
* into an AVPacket to pass data to ffmpeg without extra copies.
* This applies to buf and side_data, so this function clears them before
* freeing.
*/
void mp_free_av_packet(AVPacket **pkt)
{
if (*pkt) {
(*pkt)->side_data = NULL;
(*pkt)->side_data_elems = 0;
(*pkt)->buf = NULL;
}
av_packet_free(pkt);
}