matroska: redo seekhead handling

In particular, this reads chained seekheads. This makes seeking faster
in files which have the index indirectly linked through 2 seekheads.

As a side-effect, this warns when reading level-1 (toplevel) elements
multiple times (other than seekheads, clusters, and void/crc). Such
elements are not valid and likely break everything.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
This commit is contained in:
wm4 2015-02-09 20:39:00 +01:00 committed by Michael Niedermayer
parent a5c9befbf4
commit cac2295b21
1 changed files with 88 additions and 37 deletions

View File

@ -71,6 +71,7 @@ typedef enum {
EBML_UTF8,
EBML_BIN,
EBML_NEST,
EBML_LEVEL1,
EBML_PASS,
EBML_STOP,
EBML_SINT,
@ -252,6 +253,12 @@ typedef struct {
EbmlList blocks;
} MatroskaCluster;
typedef struct {
uint64_t id;
uint64_t pos;
int parsed;
} MatroskaLevel1Element;
typedef struct {
AVFormatContext *ctx;
@ -290,6 +297,10 @@ typedef struct {
/* File has a CUES element, but we defer parsing until it is needed. */
int cues_parsing_deferred;
/* Level1 elements and whether they were read yet */
MatroskaLevel1Element level1_elems[64];
int num_level1_elems;
int current_cluster_num_blocks;
int64_t current_cluster_pos;
MatroskaCluster current_cluster;
@ -551,13 +562,13 @@ static EbmlSyntax matroska_seekhead[] = {
};
static EbmlSyntax matroska_segment[] = {
{ MATROSKA_ID_INFO, EBML_NEST, 0, 0, { .n = matroska_info } },
{ MATROSKA_ID_TRACKS, EBML_NEST, 0, 0, { .n = matroska_tracks } },
{ MATROSKA_ID_ATTACHMENTS, EBML_NEST, 0, 0, { .n = matroska_attachments } },
{ MATROSKA_ID_CHAPTERS, EBML_NEST, 0, 0, { .n = matroska_chapters } },
{ MATROSKA_ID_CUES, EBML_NEST, 0, 0, { .n = matroska_index } },
{ MATROSKA_ID_TAGS, EBML_NEST, 0, 0, { .n = matroska_tags } },
{ MATROSKA_ID_SEEKHEAD, EBML_NEST, 0, 0, { .n = matroska_seekhead } },
{ MATROSKA_ID_INFO, EBML_LEVEL1, 0, 0, { .n = matroska_info } },
{ MATROSKA_ID_TRACKS, EBML_LEVEL1, 0, 0, { .n = matroska_tracks } },
{ MATROSKA_ID_ATTACHMENTS, EBML_LEVEL1, 0, 0, { .n = matroska_attachments } },
{ MATROSKA_ID_CHAPTERS, EBML_LEVEL1, 0, 0, { .n = matroska_chapters } },
{ MATROSKA_ID_CUES, EBML_LEVEL1, 0, 0, { .n = matroska_index } },
{ MATROSKA_ID_TAGS, EBML_LEVEL1, 0, 0, { .n = matroska_tags } },
{ MATROSKA_ID_SEEKHEAD, EBML_LEVEL1, 0, 0, { .n = matroska_seekhead } },
{ MATROSKA_ID_CLUSTER, EBML_STOP },
{ 0 }
};
@ -976,6 +987,42 @@ static int ebml_parse_nest(MatroskaDemuxContext *matroska, EbmlSyntax *syntax,
return res;
}
/*
* Allocate and return the entry for the level1 element with the given ID. If
* an entry already exists, return the existing entry.
*/
static MatroskaLevel1Element *matroska_find_level1_elem(MatroskaDemuxContext *matroska,
uint32_t id)
{
int i;
MatroskaLevel1Element *elem;
// Some files link to all clusters; useless.
if (id == MATROSKA_ID_CLUSTER)
return NULL;
// There can be multiple seekheads.
if (id != MATROSKA_ID_SEEKHEAD) {
for (i = 0; i < matroska->num_level1_elems; i++) {
if (matroska->level1_elems[i].id == id)
return &matroska->level1_elems[i];
}
}
// Only a completely broken file would have more elements.
// It also provides a low-effort way to escape from circular seekheads
// (every iteration will add a level1 entry).
if (matroska->num_level1_elems >= FF_ARRAY_ELEMS(matroska->level1_elems)) {
av_log(matroska->ctx, AV_LOG_ERROR, "Too many level1 elements or circular seekheads.\n");
return NULL;
}
elem = &matroska->level1_elems[matroska->num_level1_elems++];
*elem = (MatroskaLevel1Element){.id = id};
return elem;
}
static int ebml_parse_elem(MatroskaDemuxContext *matroska,
EbmlSyntax *syntax, void *data)
{
@ -994,6 +1041,7 @@ static int ebml_parse_elem(MatroskaDemuxContext *matroska,
uint64_t length;
int res;
void *newelem;
MatroskaLevel1Element *level1_elem;
data = (char *) data + syntax->data_offset;
if (syntax->list_elem_size) {
@ -1036,11 +1084,20 @@ static int ebml_parse_elem(MatroskaDemuxContext *matroska,
case EBML_BIN:
res = ebml_read_binary(pb, length, data);
break;
case EBML_LEVEL1:
case EBML_NEST:
if ((res = ebml_read_master(matroska, length)) < 0)
return res;
if (id == MATROSKA_ID_SEGMENT)
matroska->segment_start = avio_tell(matroska->ctx->pb);
if (id == MATROSKA_ID_CUES)
matroska->cues_parsing_deferred = 0;
if (syntax->type == EBML_LEVEL1 &&
(level1_elem = matroska_find_level1_elem(matroska, syntax->id))) {
if (level1_elem->parsed)
av_log(matroska->ctx, AV_LOG_ERROR, "Duplicate element\n");
level1_elem->parsed = 1;
}
return ebml_parse_nest(matroska, syntax->def.n, data);
case EBML_PASS:
return ebml_parse_id(matroska, syntax->def.n, id, data);
@ -1071,6 +1128,7 @@ static void ebml_free(EbmlSyntax *syntax, void *data)
case EBML_BIN:
av_freep(&((EbmlBin *) data_off)->data);
break;
case EBML_LEVEL1:
case EBML_NEST:
if (syntax[i].list_elem_size) {
EbmlList *list = data_off;
@ -1356,24 +1414,17 @@ static void matroska_convert_tags(AVFormatContext *s)
}
static int matroska_parse_seekhead_entry(MatroskaDemuxContext *matroska,
int idx)
uint64_t pos)
{
EbmlList *seekhead_list = &matroska->seekhead;
uint32_t level_up = matroska->level_up;
uint32_t saved_id = matroska->current_id;
MatroskaSeekhead *seekhead = seekhead_list->elem;
int64_t before_pos = avio_tell(matroska->ctx->pb);
MatroskaLevel level;
int64_t offset;
int ret = 0;
if (idx >= seekhead_list->nb_elem ||
seekhead[idx].id == MATROSKA_ID_SEEKHEAD ||
seekhead[idx].id == MATROSKA_ID_CLUSTER)
return 0;
/* seek */
offset = seekhead[idx].pos + matroska->segment_start;
offset = pos + matroska->segment_start;
if (avio_seek(matroska->ctx->pb, offset, SEEK_SET) == offset) {
/* We don't want to lose our seekhead level, so we add
* a dummy. This is a crude hack. */
@ -1410,37 +1461,35 @@ static int matroska_parse_seekhead_entry(MatroskaDemuxContext *matroska,
static void matroska_execute_seekhead(MatroskaDemuxContext *matroska)
{
EbmlList *seekhead_list = &matroska->seekhead;
int64_t before_pos = avio_tell(matroska->ctx->pb);
int i;
int nb_elem;
// we should not do any seeking in the streaming case
if (!matroska->ctx->pb->seekable ||
(matroska->ctx->flags & AVFMT_FLAG_IGNIDX))
return;
// do not read entries that are added while parsing seekhead entries
nb_elem = seekhead_list->nb_elem;
for (i = 0; i < seekhead_list->nb_elem; i++) {
MatroskaSeekhead *seekheads = seekhead_list->elem;
uint32_t id = seekheads[i].id;
uint64_t pos = seekheads[i].pos;
for (i = 0; i < nb_elem; i++) {
MatroskaSeekhead *seekhead = seekhead_list->elem;
if (seekhead[i].pos <= before_pos)
MatroskaLevel1Element *elem = matroska_find_level1_elem(matroska, id);
if (!elem || elem->parsed)
continue;
elem->pos = pos;
// defer cues parsing until we actually need cue data.
if (seekhead[i].id == MATROSKA_ID_CUES) {
matroska->cues_parsing_deferred = 1;
if (id == MATROSKA_ID_CUES)
continue;
}
if (matroska_parse_seekhead_entry(matroska, i) < 0) {
if (matroska_parse_seekhead_entry(matroska, pos) < 0) {
// mark index as broken
matroska->cues_parsing_deferred = -1;
break;
}
}
if (nb_elem != seekhead_list->nb_elem) {
avpriv_request_sample(matroska->ctx, "recursive SeekHead elements");
elem->parsed = 1;
}
}
@ -1474,17 +1523,18 @@ static void matroska_add_index_entries(MatroskaDemuxContext *matroska)
}
static void matroska_parse_cues(MatroskaDemuxContext *matroska) {
EbmlList *seekhead_list = &matroska->seekhead;
MatroskaSeekhead *seekhead = seekhead_list->elem;
int i;
for (i = 0; i < seekhead_list->nb_elem; i++)
if (seekhead[i].id == MATROSKA_ID_CUES)
for (i = 0; i < matroska->num_level1_elems; i++) {
MatroskaLevel1Element *elem = &matroska->level1_elems[i];
if (elem->id == MATROSKA_ID_CUES && !elem->parsed) {
if (matroska_parse_seekhead_entry(matroska, elem->pos) < 0)
matroska->cues_parsing_deferred = -1;
elem->parsed = 1;
break;
av_assert1(i <= seekhead_list->nb_elem);
}
}
if (matroska_parse_seekhead_entry(matroska, i) < 0)
matroska->cues_parsing_deferred = -1;
matroska_add_index_entries(matroska);
}
@ -2014,6 +2064,7 @@ static int matroska_read_header(AVFormatContext *s)
int i, j, res;
matroska->ctx = s;
matroska->cues_parsing_deferred = 1;
/* First read the EBML header. */
if (ebml_parse(matroska, ebml_syntax, &ebml) ||