FFmpeg/libavformat/hlsenc.c

639 lines
18 KiB
C
Raw Normal View History

/*
* Apple HTTP Live Streaming segmenter
* Copyright (c) 2012, Luca Barbato
*
* This file is part of Libav.
*
* Libav 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.
*
* Libav 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 Libav; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include <float.h>
#include <stdint.h>
#include <config.h>
#if CONFIG_OPENSSL
#include <openssl/rand.h>
#endif
#include "libavutil/mathematics.h"
#include "libavutil/parseutils.h"
#include "libavutil/avstring.h"
#include "libavutil/intreadwrite.h"
#include "libavutil/opt.h"
#include "libavutil/random_seed.h"
#include "libavutil/log.h"
#include "avformat.h"
#include "internal.h"
typedef struct ListEntry {
char name[1024];
int64_t duration; // segment duration in AV_TIME_BASE units
int discont;
struct ListEntry *next;
} ListEntry;
typedef struct HLSContext {
const AVClass *class; // Class for private options.
unsigned number;
int64_t sequence;
int64_t start_sequence;
AVOutputFormat *oformat;
AVFormatContext *avf;
float time; // Set by a private option.
int size; // Set by a private option.
int wrap; // Set by a private option.
int version; // Set by a private option.
int allowcache;
int64_t recording_time;
int has_video;
// The following timestamps are in AV_TIME_BASE units.
int64_t start_pts;
int64_t end_pts;
int64_t duration; // last segment duration computed so far.
int nb_entries;
ListEntry *list;
ListEntry *end_list;
char *basename;
char *baseurl;
int encrypt; // Set by a private option.
char *key; // Set by a private option.
int key_len;
char *key_url; // Set by a private option.
char *iv; // Set by a private option.
int iv_len;
int recovered;
char *key_basename;
AVDictionary *enc_opts;
} HLSContext;
static int randomize(uint8_t *buf, int len)
{
#if CONFIG_OPENSSL
if (RAND_bytes(buf, len))
return 0;
2016-11-18 09:56:01 +00:00
return AVERROR(EIO);
#else
return AVERROR(ENOSYS);
#endif
}
static void free_encryption(AVFormatContext *s)
{
HLSContext *hls = s->priv_data;
av_dict_free(&hls->enc_opts);
av_freep(&hls->key_basename);
}
static int dict_set_bin(AVDictionary **dict, const char *key,
uint8_t *buf, size_t len)
{
char hex[33];
ff_data_to_hex(hex, buf, len, 0);
hex[32] = '\0';
return av_dict_set(dict, key, hex, 0);
}
static int setup_encryption(AVFormatContext *s)
{
HLSContext *hls = s->priv_data;
AVIOContext *out = NULL;
int len, ret;
uint8_t buf[16];
uint8_t *k = NULL;
len = strlen(hls->basename) + 4 + 1;
hls->key_basename = av_mallocz(len);
if (!hls->key_basename)
return AVERROR(ENOMEM);
av_strlcpy(hls->key_basename, hls->basename + 7, len);
av_strlcat(hls->key_basename, ".key", len);
if (hls->key) {
if (hls->key_len != 16) {
av_log(s, AV_LOG_ERROR,
"Invalid key size %d, expected 16-bytes hex-coded key\n",
hls->key_len);
return AVERROR(EINVAL);
}
if ((ret = dict_set_bin(&hls->enc_opts, "key", hls->key, hls->key_len)) < 0)
return ret;
k = hls->key;
} else {
if (hls->start_sequence < 0) {
ret = s->io_open(s, &out, hls->key_basename, AVIO_FLAG_READ, NULL);
if (ret < 0) {
av_log(s, AV_LOG_WARNING,
"Cannot recover the key, generating a new one.\n");
} else {
avio_read(out, buf, 16);
k = buf;
avio_close(out);
}
}
if (!k) {
if ((ret = randomize(buf, sizeof(buf))) < 0) {
av_log(s, AV_LOG_ERROR, "Cannot generate a strong random key\n");
return ret;
}
}
if ((ret = dict_set_bin(&hls->enc_opts, "key", buf, sizeof(buf))) < 0)
return ret;
k = buf;
}
if (hls->iv) {
if (hls->iv_len != 16) {
av_log(s, AV_LOG_ERROR,
"Invalid key size %d, expected 16-bytes hex-coded initialization vector\n",
hls->iv_len);
return AVERROR(EINVAL);
}
if ((ret = dict_set_bin(&hls->enc_opts, "iv", hls->iv, hls->iv_len)) < 0)
return ret;
}
if ((ret = s->io_open(s, &out, hls->key_basename, AVIO_FLAG_WRITE, NULL)) < 0)
return ret;
avio_write(out, k, 16);
avio_close(out);
return 0;
}
static int hls_mux_init(AVFormatContext *s)
{
HLSContext *hls = s->priv_data;
AVFormatContext *oc;
int i;
hls->avf = oc = avformat_alloc_context();
if (!oc)
return AVERROR(ENOMEM);
oc->oformat = hls->oformat;
oc->interrupt_callback = s->interrupt_callback;
oc->opaque = s->opaque;
oc->io_open = s->io_open;
oc->io_close = s->io_close;
for (i = 0; i < s->nb_streams; i++) {
AVStream *st;
if (!(st = avformat_new_stream(oc, NULL)))
return AVERROR(ENOMEM);
lavf: replace AVStream.codec with AVStream.codecpar Currently, AVStream contains an embedded AVCodecContext instance, which is used by demuxers to export stream parameters to the caller and by muxers to receive stream parameters from the caller. It is also used internally as the codec context that is passed to parsers. In addition, it is also widely used by the callers as the decoding (when demuxer) or encoding (when muxing) context, though this has been officially discouraged since Libav 11. There are multiple important problems with this approach: - the fields in AVCodecContext are in general one of * stream parameters * codec options * codec state However, it's not clear which ones are which. It is consequently unclear which fields are a demuxer allowed to set or a muxer allowed to read. This leads to erratic behaviour depending on whether decoding or encoding is being performed or not (and whether it uses the AVStream embedded codec context). - various synchronization issues arising from the fact that the same context is used by several different APIs (muxers/demuxers, parsers, bitstream filters and encoders/decoders) simultaneously, with there being no clear rules for who can modify what and the different processes being typically delayed with respect to each other. - avformat_find_stream_info() making it necessary to support opening and closing a single codec context multiple times, thus complicating the semantics of freeing various allocated objects in the codec context. Those problems are resolved by replacing the AVStream embedded codec context with a newly added AVCodecParameters instance, which stores only the stream parameters exported by the demuxers or read by the muxers.
2014-06-18 18:42:52 +00:00
avcodec_parameters_copy(st->codecpar, s->streams[i]->codecpar);
st->sample_aspect_ratio = s->streams[i]->sample_aspect_ratio;
st->time_base = s->streams[i]->time_base;
}
return 0;
}
static int append_entry(HLSContext *hls, int64_t duration, const char *name,
int discont)
{
ListEntry *en = av_malloc(sizeof(*en));
if (!en)
return AVERROR(ENOMEM);
av_strlcpy(en->name, name, sizeof(en->name));
en->discont = discont;
en->duration = duration;
en->next = NULL;
if (!hls->list)
hls->list = en;
else
hls->end_list->next = en;
hls->end_list = en;
if (hls->nb_entries >= hls->size) {
en = hls->list;
hls->list = en->next;
av_free(en);
} else
hls->nb_entries++;
hls->sequence++;
return 0;
}
static void free_entries(HLSContext *hls)
{
ListEntry *p = hls->list, *en;
while(p) {
en = p;
p = p->next;
av_free(en);
}
}
static int hls_window(AVFormatContext *s, int last)
{
HLSContext *hls = s->priv_data;
ListEntry *en;
int64_t target_duration = 0;
int ret = 0;
AVIOContext *out = NULL;
char temp_filename[1024];
2014-04-29 16:37:01 +00:00
int64_t sequence = FFMAX(hls->start_sequence, hls->sequence - hls->size);
snprintf(temp_filename, sizeof(temp_filename), "%s.tmp", s->filename);
if ((ret = s->io_open(s, &out, temp_filename, AVIO_FLAG_WRITE, NULL)) < 0)
goto fail;
for (en = hls->list; en; en = en->next) {
if (target_duration < en->duration)
target_duration = en->duration;
}
avio_printf(out, "#EXTM3U\n");
avio_printf(out, "#EXT-X-VERSION:%d\n", hls->version);
if (hls->allowcache == 0 || hls->allowcache == 1) {
avio_printf(out, "#EXT-X-ALLOW-CACHE:%s\n", hls->allowcache == 0 ? "NO" : "YES");
}
avio_printf(out, "#EXT-X-TARGETDURATION:%"PRId64"\n",
av_rescale_rnd(target_duration, 1, AV_TIME_BASE,
AV_ROUND_UP));
avio_printf(out, "#EXT-X-MEDIA-SEQUENCE:%"PRId64"\n", sequence);
av_log(s, AV_LOG_VERBOSE, "EXT-X-MEDIA-SEQUENCE:%"PRId64"\n",
2014-04-29 16:37:01 +00:00
sequence);
for (en = hls->list; en; en = en->next) {
if (en->discont) {
avio_printf(out, "#EXT-X-DISCONTINUITY\n");
}
if (hls->encrypt) {
char *key_url;
if (hls->key_url)
key_url = hls->key_url;
else
key_url = hls->baseurl;
avio_printf(out, "#EXT-X-KEY:METHOD=AES-128");
avio_printf(out, ",URI=\"");
if (key_url)
avio_printf(out, "%s", key_url);
avio_printf(out, "%s\"", av_basename(hls->key_basename));
if (hls->iv)
avio_printf(out, ",IV=\"0x%s\"", hls->iv);
avio_printf(out, "\n");
}
if (hls->version > 2)
avio_printf(out, "#EXTINF:%f\n",
(double)en->duration / AV_TIME_BASE);
else
avio_printf(out, "#EXTINF:%"PRId64",\n",
av_rescale(en->duration, 1, AV_TIME_BASE));
if (hls->baseurl)
avio_printf(out, "%s", hls->baseurl);
avio_printf(out, "%s\n", en->name);
}
if (last)
avio_printf(out, "#EXT-X-ENDLIST\n");
fail:
ff_format_io_close(s, &out);
if (ret >= 0)
ff_rename(temp_filename, s->filename);
return ret;
}
static int hls_start(AVFormatContext *s)
{
HLSContext *c = s->priv_data;
AVFormatContext *oc = c->avf;
int err = 0;
AVDictionary *opts = NULL;
if (av_get_frame_filename(oc->filename, sizeof(oc->filename),
c->basename, c->wrap ? c->sequence % c->wrap : c->sequence) < 0)
return AVERROR(EINVAL);
c->number++;
if (c->encrypt) {
if ((err = av_dict_copy(&opts, c->enc_opts, 0)) < 0)
return err;
if (!c->iv) {
uint8_t iv[16] = { 0 };
char buf[33];
AV_WB64(iv + 8, c->sequence);
ff_data_to_hex(buf, iv, sizeof(iv), 0);
buf[32] = '\0';
if ((err = av_dict_set(&opts, "iv", buf, 0)) < 0)
goto fail;
}
}
if ((err = s->io_open(s, &oc->pb, oc->filename, AVIO_FLAG_WRITE, &opts)) < 0)
return err;
if (oc->oformat->priv_class && oc->priv_data)
av_opt_set(oc->priv_data, "mpegts_flags", "resend_headers", 0);
fail:
av_dict_free(&opts);
return err;
}
static int read_chomp_line(AVIOContext *s, char *buf, int maxlen)
{
int len = ff_get_line(s, buf, maxlen);
while (len > 0 && av_isspace(buf[len - 1]))
buf[--len] = '\0';
return len;
}
static int hls_recover(AVFormatContext *s)
{
HLSContext *hls = s->priv_data;
char line[1024];
AVIOContext *io;
const char *ptr;
int ret, is_segment = 0, is_discont = 0;
int64_t duration = 0;
ret = s->io_open(s, &io, s->filename, AVIO_FLAG_READ, NULL);
if (ret < 0) {
av_log(s, AV_LOG_WARNING,
"Cannot recover the playlist, generating a new one.\n");
hls->start_sequence = 0;
hls->sequence = 0;
return 0;
}
read_chomp_line(io, line, sizeof(line));
if (strcmp(line, "#EXTM3U")) {
av_log(s, AV_LOG_ERROR,
"The playlist file is present but unparsable."
" Please remove it.\n");
return AVERROR_INVALIDDATA;
}
while (!io->eof_reached) {
read_chomp_line(io, line, sizeof(line));
if (av_strstart(line, "#EXT-X-MEDIA-SEQUENCE:", &ptr)) {
hls->sequence = hls->start_sequence = atoi(ptr);
} else if (av_strstart(line, "#EXTINF:", &ptr)) {
is_segment = 1;
duration = atof(ptr) * AV_TIME_BASE;
} else if (av_strstart(line, "#EXT-X-DISCONTINUITY", NULL)) {
is_discont = 1;
} else if (av_strstart(line, "#", NULL)) {
continue;
} else if (line[0]) {
if (is_segment) {
append_entry(hls, duration, av_basename(line), is_discont);
is_segment = 0;
is_discont = 0;
}
}
}
hls->recovered = 1;
return 0;
}
static int hls_setup(AVFormatContext *s)
{
HLSContext *hls = s->priv_data;
const char *pattern = "%d.ts";
int basename_size = strlen(s->filename) + strlen(pattern) + 1;
char *p;
int ret;
if (hls->encrypt)
basename_size += 7;
hls->basename = av_mallocz(basename_size);
if (!hls->basename)
return AVERROR(ENOMEM);
// TODO: support protocol nesting?
if (hls->encrypt)
strcpy(hls->basename, "crypto:");
av_strlcat(hls->basename, s->filename, basename_size);
p = strrchr(hls->basename, '.');
if (p)
*p = '\0';
if (hls->encrypt) {
ret = setup_encryption(s);
if (ret < 0)
return ret;
}
if (hls->start_sequence < 0) {
ret = hls_recover(s);
if (ret < 0)
return ret;
}
av_strlcat(hls->basename, pattern, basename_size);
return 0;
}
static int hls_write_header(AVFormatContext *s)
{
HLSContext *hls = s->priv_data;
int ret, i;
hls->sequence = hls->start_sequence;
hls->recording_time = hls->time * AV_TIME_BASE;
hls->start_pts = AV_NOPTS_VALUE;
for (i = 0; i < s->nb_streams; i++)
hls->has_video +=
lavf: replace AVStream.codec with AVStream.codecpar Currently, AVStream contains an embedded AVCodecContext instance, which is used by demuxers to export stream parameters to the caller and by muxers to receive stream parameters from the caller. It is also used internally as the codec context that is passed to parsers. In addition, it is also widely used by the callers as the decoding (when demuxer) or encoding (when muxing) context, though this has been officially discouraged since Libav 11. There are multiple important problems with this approach: - the fields in AVCodecContext are in general one of * stream parameters * codec options * codec state However, it's not clear which ones are which. It is consequently unclear which fields are a demuxer allowed to set or a muxer allowed to read. This leads to erratic behaviour depending on whether decoding or encoding is being performed or not (and whether it uses the AVStream embedded codec context). - various synchronization issues arising from the fact that the same context is used by several different APIs (muxers/demuxers, parsers, bitstream filters and encoders/decoders) simultaneously, with there being no clear rules for who can modify what and the different processes being typically delayed with respect to each other. - avformat_find_stream_info() making it necessary to support opening and closing a single codec context multiple times, thus complicating the semantics of freeing various allocated objects in the codec context. Those problems are resolved by replacing the AVStream embedded codec context with a newly added AVCodecParameters instance, which stores only the stream parameters exported by the demuxers or read by the muxers.
2014-06-18 18:42:52 +00:00
s->streams[i]->codecpar->codec_type == AVMEDIA_TYPE_VIDEO;
if (hls->has_video > 1)
av_log(s, AV_LOG_WARNING,
"More than a single video stream present, "
"expect issues decoding it.\n");
hls->oformat = av_guess_format("mpegts", NULL, NULL);
if (!hls->oformat) {
ret = AVERROR_MUXER_NOT_FOUND;
goto fail;
}
if ((ret = hls_setup(s)) < 0)
goto fail;
if ((ret = hls_mux_init(s)) < 0)
goto fail;
if ((ret = hls_start(s)) < 0)
goto fail;
if ((ret = avformat_write_header(hls->avf, NULL)) < 0)
return ret;
fail:
if (ret) {
av_free(hls->basename);
if (hls->avf)
avformat_free_context(hls->avf);
free_encryption(s);
}
return ret;
}
static int hls_write_packet(AVFormatContext *s, AVPacket *pkt)
{
HLSContext *hls = s->priv_data;
AVFormatContext *oc = hls->avf;
AVStream *st = s->streams[pkt->stream_index];
int64_t end_pts = hls->recording_time * hls->number;
int64_t pts = av_rescale_q(pkt->pts, st->time_base, AV_TIME_BASE_Q);
int ret, can_split = 1;
if (hls->start_pts == AV_NOPTS_VALUE) {
hls->start_pts = pts;
hls->end_pts = pts;
}
if (hls->has_video) {
lavf: replace AVStream.codec with AVStream.codecpar Currently, AVStream contains an embedded AVCodecContext instance, which is used by demuxers to export stream parameters to the caller and by muxers to receive stream parameters from the caller. It is also used internally as the codec context that is passed to parsers. In addition, it is also widely used by the callers as the decoding (when demuxer) or encoding (when muxing) context, though this has been officially discouraged since Libav 11. There are multiple important problems with this approach: - the fields in AVCodecContext are in general one of * stream parameters * codec options * codec state However, it's not clear which ones are which. It is consequently unclear which fields are a demuxer allowed to set or a muxer allowed to read. This leads to erratic behaviour depending on whether decoding or encoding is being performed or not (and whether it uses the AVStream embedded codec context). - various synchronization issues arising from the fact that the same context is used by several different APIs (muxers/demuxers, parsers, bitstream filters and encoders/decoders) simultaneously, with there being no clear rules for who can modify what and the different processes being typically delayed with respect to each other. - avformat_find_stream_info() making it necessary to support opening and closing a single codec context multiple times, thus complicating the semantics of freeing various allocated objects in the codec context. Those problems are resolved by replacing the AVStream embedded codec context with a newly added AVCodecParameters instance, which stores only the stream parameters exported by the demuxers or read by the muxers.
2014-06-18 18:42:52 +00:00
can_split = st->codecpar->codec_type == AVMEDIA_TYPE_VIDEO &&
pkt->flags & AV_PKT_FLAG_KEY;
}
if (pkt->pts == AV_NOPTS_VALUE)
can_split = 0;
else
hls->duration = pts - hls->end_pts;
if (can_split && pts - hls->start_pts >= end_pts) {
ret = append_entry(hls, hls->duration, av_basename(hls->avf->filename), hls->recovered);
hls->recovered = 0;
if (ret)
return ret;
hls->end_pts = pts;
hls->duration = 0;
av_write_frame(oc, NULL); /* Flush any buffered data */
ff_format_io_close(s, &oc->pb);
ret = hls_start(s);
if (ret)
return ret;
oc = hls->avf;
if ((ret = hls_window(s, 0)) < 0)
return ret;
}
ret = ff_write_chained(oc, pkt->stream_index, pkt, s);
return ret;
}
static int hls_write_trailer(struct AVFormatContext *s)
{
HLSContext *hls = s->priv_data;
AVFormatContext *oc = hls->avf;
av_write_trailer(oc);
ff_format_io_close(s, &oc->pb);
avformat_free_context(oc);
av_free(hls->basename);
append_entry(hls, hls->duration, av_basename(hls->avf->filename), 0);
hls_window(s, 1);
free_entries(hls);
free_encryption(s);
return 0;
}
#define OFFSET(x) offsetof(HLSContext, x)
#define E AV_OPT_FLAG_ENCODING_PARAM
static const AVOption options[] = {
{"start_number", "first number in the sequence", OFFSET(start_sequence),AV_OPT_TYPE_INT64, {.i64 = 0}, -1, INT64_MAX, E, "start_number"},
{"recover", "If there is already a m3u8 file in the path, populate the sequence from it", 0, AV_OPT_TYPE_CONST, {.i64 = -1}, 0, 0, E, "start_number"},
{"hls_time", "segment length in seconds", OFFSET(time), AV_OPT_TYPE_FLOAT, {.dbl = 2}, 0, FLT_MAX, E},
{"hls_list_size", "maximum number of playlist entries", OFFSET(size), AV_OPT_TYPE_INT, {.i64 = 5}, 0, INT_MAX, E},
{"hls_wrap", "number after which the index wraps", OFFSET(wrap), AV_OPT_TYPE_INT, {.i64 = 0}, 0, INT_MAX, E},
{"hls_allow_cache", "explicitly set whether the client MAY (1) or MUST NOT (0) cache media segments", OFFSET(allowcache), AV_OPT_TYPE_INT, {.i64 = -1}, INT_MIN, INT_MAX, E},
{"hls_base_url", "url to prepend to each playlist entry", OFFSET(baseurl), AV_OPT_TYPE_STRING, {.str = NULL}, 0, 0, E},
{"hls_version", "protocol version", OFFSET(version), AV_OPT_TYPE_INT, {.i64 = 3}, 2, 3, E},
{"hls_enc", "AES128 encryption support", OFFSET(encrypt), AV_OPT_TYPE_INT, {.i64 = 0}, 0, 1, E},
{"hls_enc_key", "use the specified hex-coded 16byte key to encrypt the segments", OFFSET(key), AV_OPT_TYPE_BINARY, .flags = E},
{"hls_enc_key_url", "url to access the key to decrypt the segments", OFFSET(key_url), AV_OPT_TYPE_STRING, {.str = NULL}, 0, 0, E},
{"hls_enc_iv", "use the specified hex-coded 16byte initialization vector", OFFSET(iv), AV_OPT_TYPE_BINARY, .flags = E},
{ NULL },
};
static const AVClass hls_class = {
.class_name = "hls muxer",
.item_name = av_default_item_name,
.option = options,
.version = LIBAVUTIL_VERSION_INT,
};
AVOutputFormat ff_hls_muxer = {
.name = "hls",
2012-12-08 05:00:46 +00:00
.long_name = NULL_IF_CONFIG_SMALL("Apple HTTP Live Streaming"),
.extensions = "m3u8",
.priv_data_size = sizeof(HLSContext),
.audio_codec = AV_CODEC_ID_AAC,
.video_codec = AV_CODEC_ID_H264,
.flags = AVFMT_NOFILE | AVFMT_ALLOW_FLUSH,
.write_header = hls_write_header,
.write_packet = hls_write_packet,
.write_trailer = hls_write_trailer,
.priv_class = &hls_class,
};