[asterisk-dev] WebRTC SFU: support add video track dynamically

[Xiemin Chen]

Please see the comments in the following codes, at the moment, A & B both
have one audio & video stream separately and B is starting to add an extra
sharing video:

static int sdp_requires_deferral(struct ast_sip_session *session, const
pjmedia_sdp_session *sdp)
{
int i;

if (!session->pending_media_state->topology) {
session->pending_media_state->topology = ast_stream_topology_alloc();
if (!session->pending_media_state->topology) {
return -1;
}
}

for (i = 0; i < sdp->media_count; ++i) {
/* See if there are registered handlers for this media stream type */
char media[20];
struct ast_sip_session_sdp_handler *handler;
RAII_VAR(struct sdp_handler_list *, handler_list, NULL, ao2_cleanup);
struct ast_stream *existing_stream = NULL;
struct ast_stream *stream;
enum ast_media_type type;
struct ast_sip_session_media *session_media = NULL;
enum ast_sip_session_sdp_stream_defer res;

/* We need a null-terminated version of the media string */
ast_copy_pj_str(media, &sdp->media[i]->desc.media, sizeof(media));

                /* The stream count of B's active_media_state->topology is
3, one B's audio, one B's video, one A's video with name
softbridge_dest_PJSIP...
                   The sdp->media_count is 3, one for audio, one for camera
video, one for sharing video
                   Here B's sharing video is map to A's video with name
softbridge_dest_PJSIP...  */
if (session->active_media_state->topology &&
(i < ast_stream_topology_get_count(session->active_media_state->topology)))
{
existing_stream =
ast_stream_topology_get_stream(session->active_media_state->topology, i);
}

type = ast_media_type_from_str(media);
stream = ast_stream_alloc(existing_stream ?
ast_stream_get_name(existing_stream) : ast_codec_media_type2str(type),
type);
if (!stream) {
return -1;
}

/* As this is only called on an incoming SDP offer before processing it is
not possible
* for streams and their media sessions to exist.
*/
if (ast_stream_topology_set_stream(session->pending_media_state->topology,
i, stream)) {
ast_stream_free(stream);
return -1;
}

session_media = ast_sip_session_media_state_add(session,
session->pending_media_state, ast_media_type_from_str(media), i);
if (!session_media) {
return -1;
}

               /* If the code of B's new sharing video stream executes
here, does it need to call set_mid_and_bundle_group()
& set_remote_mslabel_and_stream_group()
                  &  handler->negotiate_incoming_sdp_stream for the new
stream setup instead of the handler->defer_incoming_sdp_stream()?
               */

if (session_media->handler) {
handler = session_media->handler;
if (handler->defer_incoming_sdp_stream) {
res = handler->defer_incoming_sdp_stream(session, session_media, sdp,
sdp->media[i]);
switch (res) {
case AST_SIP_SESSION_SDP_DEFER_NOT_HANDLED:
break;
case AST_SIP_SESSION_SDP_DEFER_ERROR:
return 0;
case AST_SIP_SESSION_SDP_DEFER_NOT_NEEDED:
break;
case AST_SIP_SESSION_SDP_DEFER_NEEDED:
return 1;
}
}
/* Handled by this handler. Move to the next stream */
continue;
}

handler_list = ao2_find(sdp_handlers, media, OBJ_KEY);
if (!handler_list) {
ast_debug(1, "No registered SDP handlers for media type '%s'\n", media);
continue;
}
AST_LIST_TRAVERSE(&handler_list->list, handler, next) {
if (handler == session_media->handler) {
continue;
}
if (!handler->defer_incoming_sdp_stream) {
continue;
}
res = handler->defer_incoming_sdp_stream(session, session_media, sdp,
sdp->media[i]);
switch (res) {
case AST_SIP_SESSION_SDP_DEFER_NOT_HANDLED:
continue;
case AST_SIP_SESSION_SDP_DEFER_ERROR:
session_media_set_handler(session_media, handler);
return 0;
case AST_SIP_SESSION_SDP_DEFER_NOT_NEEDED:
/* Handled by this handler. */
session_media_set_handler(session_media, handler);
break;
case AST_SIP_SESSION_SDP_DEFER_NEEDED:
/* Handled by this handler. */
session_media_set_handler(session_media, handler);
return 1;
}
/* Move to the next stream */
break;
}
}
return 0;
}


Joshua C. Colp <jcolp at digium.com> 于2019年5月8日周三 下午5:14写道：

> On Wed, May 8, 2019, at 12:07 AM, Xiemin Chen wrote:
> > Hi there,
> >
> > *Problem*
> >
> > In a WebRTC conference, participant A & B join with only one video
> > track, then B add another video track with the SDP renegotiate, A
> > cannot see B's two video tracks.
> >
> > *Root Cause*
> > *
> > *
> > After work through the source code, the problem may caused by the
> > sdp_requires_deferral() in res_pjsip_session.c:
> > It does not take of the situation of the new creation of the video
> > track.
> >
> > *Solution*
> >
> > Can I use the handle_incoming_sdp() to replace the
> > sdp_requires_deferal() ? handle_incoming_sdp() function is called
> > during the init of the session, it does the track initial progress.
>
> The sdp_requires_deferral function and its functionality has to continue
> to exist, or T.38 fax will be non-functional. The sdp_requires_deferral
> function would need to be examined based on your scenario and adjusted
> accordingly however it makes sense. Looking at the function I'm also not
> sure why it would cause any problems with your scenario, as it should just
> result in some things being set up and that's pretty much it but I have not
> looked into it deeply or reproduced the problem.
>
> --
> Joshua C. Colp
> Digium - A Sangoma Company | Senior Software Developer
> 445 Jan Davis Drive NW - Huntsville, AL 35806 - US
> Check us out at: www.digium.com & www.asterisk.org
>
> --
> _____________________________________________________________________
> -- Bandwidth and Colocation Provided by http://www.api-digital.com --
>
> asterisk-dev mailing list
> To UNSUBSCRIBE or update options visit:
>    http://lists.digium.com/mailman/listinfo/asterisk-dev
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.digium.com/pipermail/asterisk-dev/attachments/20190509/5df72ffa/attachment-0001.html>