Skip to content
GitLab
Commit def5a4cf authored by Joseph Wallace's avatar Joseph Wallace
Browse files

Implement EBML-aware parser. 

* Loop over elements in input buffer.
* Most are literally copied with their contents without
  inspection into the header or data buffers as appropriate.
* Some only copy the element header, to allow inspecting
  children elements.
* Cluster elements are identified and used as sync points.

No probing is done for keyframes *yet*
parent 50c4984c
Hide whitespace changes
Inline Side-by-side
src/format_ebml.c
View file @ def5a4cf
......@@ -47,6 +47,14 @@ typedef enum ebml_read_mode {
EBML_STATE_READING_CLUSTERS
} ebml_read_mode;
typedef enum ebml_parsing_state {
EBML_STATE_PARSING_HEADER = 0,
EBML_STATE_COPYING_TO_HEADER,
EBML_STATE_START_CLUSTER,
EBML_STATE_PARSING_CLUSTERS,
EBML_STATE_COPYING_TO_DATA
} ebml_parsing_state;
typedef enum ebml_chunk_type {
EBML_CHUNK_HEADER = 0,
EBML_CHUNK_CLUSTER_START,
......@@ -65,8 +73,9 @@ struct ebml_client_data_st {
struct ebml_st {
ebml_read_mode output_state;
ebml_parsing_state parse_state;
unsigned long long copy_len;
char *cluster_id;
int cluster_start;
int position;
......@@ -333,11 +342,9 @@ static ebml_t *ebml_create()
ebml->output_state = EBML_STATE_READING_HEADER;
ebml->header = calloc(1, EBML_HEADER_MAX_SIZE);
ebml->buffer = calloc(1, EBML_SLICE_SIZE * 4);
ebml->buffer = calloc(1, EBML_SLICE_SIZE);
ebml->input_buffer = calloc(1, EBML_SLICE_SIZE);
ebml->cluster_id = "\x1F\x43\xB6\x75";
ebml->cluster_start = -1;
return ebml;
......@@ -487,65 +494,154 @@ static unsigned char *ebml_get_write_buffer(ebml_t *ebml, int *bytes)
*/
static int ebml_wrote(ebml_t *ebml, int len)
{
int processing = 1;
int cursor = 0;
int to_copy;
unsigned char *end_of_buffer;
int b;
if (ebml->header_size == 0) {
/* Still reading header */
if ((ebml->header_position + len) > EBML_HEADER_MAX_SIZE) {
ICECAST_LOG_ERROR("EBML Header too large, failing");
return -1;
}
/*
ICECAST_LOG_DEBUG("EBML: Adding to header, ofset is %d size is %d adding %d",
ebml->header_size, ebml->header_position, len);
*/
memcpy(ebml->header + ebml->header_position, ebml->input_buffer, len);
ebml->header_position += len;
} else {
/* Header's already been determined, read into data buffer */
memcpy(ebml->buffer + ebml->position, ebml->input_buffer, len);
}
for (b = 0; b <= len - 4; b++)
{
/* Scan for cluster start marker.
* False positives are possible, but unlikely, and only
* permanently corrupt a stream if they occur while scanning
* the initial header. Else, a client can reconnect in a few
* seconds and get a real sync point.
*/
if (!memcmp(ebml->input_buffer + b, ebml->cluster_id, 4))
{
/*
ICECAST_LOG_DEBUG("EBML: found cluster");
*/
int tag_length;
unsigned long long payload_length;
int copy_state;
if (ebml->header_size == 0)
{
/* We were looking for the header end; now we've found it */
ebml->header_size = ebml->header_position - len + b;
char *segment_id = "\x18\x53\x80\x67";
char *cluster_id = "\x1F\x43\xB6\x75";
ebml->input_position += len;
end_of_buffer = ebml->input_buffer + ebml->input_position;
while (processing) {
/*ICECAST_LOG_DEBUG("Parse State: %i", ebml->parse_state);*/
switch (ebml->parse_state) {
case EBML_STATE_PARSING_HEADER:
case EBML_STATE_PARSING_CLUSTERS:
/* Shift data after the header into the data buffer */
memcpy(ebml->buffer, ebml->input_buffer + b, len - b);
ebml->position = len - b;
if (ebml->parse_state == EBML_STATE_PARSING_HEADER) {
copy_state = EBML_STATE_COPYING_TO_HEADER;
} else {
copy_state = EBML_STATE_COPYING_TO_DATA;
}
/* Mark the start of the data as the first sync point */
ebml->cluster_start = 0;
return len;
} else {
/* We've located a sync point in the data stream */
ebml->cluster_start = ebml->position + b;
}
tag_length = ebml_parse_tag(ebml->input_buffer + cursor,
end_of_buffer, &payload_length);
if (tag_length > 0) {
if (payload_length == EBML_UNKNOWN) {
/* Parse all children for tags we can't skip */
payload_length = 0;
}
if (!memcmp(ebml->input_buffer + cursor, cluster_id, 4)) {
/* Found a cluster */
ebml->parse_state = EBML_STATE_START_CLUSTER;
} else if (!memcmp(ebml->input_buffer + cursor, segment_id, 4)) {
/* Segment tag, copy tag to buffer but parse children */
ebml->copy_len = tag_length;
ebml->parse_state = copy_state;
} else {
/* Non-cluster tag, copy it & children into buffer */
ebml->copy_len = tag_length + payload_length;
ebml->parse_state = copy_state;
}
} else if (tag_length == 0) {
/* Wait for more data */
/* ICECAST_LOG_DEBUG("Wait"); */
processing = 0;
} else if (tag_length < 0) {
/* Parse error */
/* ICECAST_LOG_DEBUG("Stop"); */
return -1;
}
break;
case EBML_STATE_START_CLUSTER:
/* found a cluster; wait to process it until
* any previous cluster tag has been flushed
* from the read buffer, so as to not lose the
* sync point.
*/
if (ebml->cluster_start >= 0) {
processing = 0;
} else {
tag_length = ebml_parse_tag(ebml->input_buffer + cursor,
end_of_buffer, &payload_length);
/* The header has been fully read by now, publish its size. */
ebml->header_size = ebml->header_position;
/* Mark this sync point */
ebml->cluster_start = ebml->position;
/* Copy cluster tag to read buffer */
ebml->copy_len = tag_length;
ebml->parse_state = EBML_STATE_COPYING_TO_DATA;
}
break;
case EBML_STATE_COPYING_TO_HEADER:
case EBML_STATE_COPYING_TO_DATA:
to_copy = ebml->input_position - cursor;
if (to_copy > ebml->copy_len) {
to_copy = ebml->copy_len;
}
if (ebml->parse_state == EBML_STATE_COPYING_TO_HEADER) {
if ((ebml->header_position + to_copy) > EBML_HEADER_MAX_SIZE) {
ICECAST_LOG_ERROR("EBML Header too large, failing");
return -1;
}
memcpy(ebml->header + ebml->header_position, ebml->input_buffer + cursor, to_copy);
ebml->header_position += to_copy;
} else if (ebml->parse_state == EBML_STATE_COPYING_TO_DATA) {
if ((ebml->position + to_copy) > EBML_SLICE_SIZE) {
to_copy = EBML_SLICE_SIZE - ebml->position;
}
memcpy(ebml->buffer + ebml->position, ebml->input_buffer + cursor, to_copy);
ebml->position += to_copy;
}
/* ICECAST_LOG_DEBUG("Copied %i of %hhu", to_copy, ebml->copy_len); */
cursor += to_copy;
ebml->copy_len -= to_copy;
if (ebml->copy_len == 0) {
/* resume parsing */
if (ebml->parse_state == EBML_STATE_COPYING_TO_HEADER) {
ebml->parse_state = EBML_STATE_PARSING_HEADER;
} else {
ebml->parse_state = EBML_STATE_PARSING_CLUSTERS;
}
} else {
/* wait for more data */
processing = 0;
}
break;
default:
processing = 0;
}
}
ebml->position += len;
/* Shift unprocessed data down to the start of the buffer */
memmove(ebml->input_buffer, ebml->input_buffer + cursor, ebml->input_position - cursor);
ebml->input_position -= cursor;
return len;
}
/* Try to parse an EBML tag at the given location, returning the
......
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment