Add pass through support for Opus and VP8; Opus format attribute negotiation

This patch adds pass through support for Opus and VP8. That includes:

* Format attribute negotiation for Opus. Note that unlike some other codecs,
  the draft RFC specifies having spaces delimiting the attributes in addition
  to ';', so you have "attra=X; attrb=Y". This broke the attribute parsing in
  chan_sip, so a small tweak was also included in this patch for that.

* A format attribute negotiation module for Opus, res_format_attr_opus

* Fast picture update for VP8. Since VP8 uses a different RTCP packet number
  than FIR, this really is specific to VP8 at this time.

Note that the format attribute negotiation in res_pjsip_sdp_rtp was written
by mjordan. The rest of this patch was written completely by Lorenzo Miniero.

Review: https://reviewboard.asterisk.org/r/2723/

(closes issue ASTERISK-21981)
Reported by: Tzafrir Cohen
patches:
  asterisk_opus+vp8_passthrough_20130718.patch uploaded by lminiero (License 6518)



git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@397526 65c4cc65-6c06-0410-ace0-fbb531ad65f3

3 files changed, 385 insertions, 4 deletions

diff --git a/res/res_format_attr_opus.c b/res/res_format_attr_opus.c
new file mode 100644
index 000000000..ed8adb77f
--- /dev/null
+++ b/res/res_format_attr_opus.c
@@ -0,0 +1,321 @@
+/*
+ * Asterisk -- An open source telephony toolkit.
+ *
+ * Copyright (C) 2013, Digium, Inc.
+ *
+ * Lorenzo Miniero <lorenzo@meetecho.com>
+ *
+ * See http://www.asterisk.org for more information about
+ * the Asterisk project. Please do not directly contact
+ * any of the maintainers of this project for assistance;
+ * the project provides a web site, mailing lists and IRC
+ * channels for your use.
+ *
+ * This program is free software, distributed under the terms of
+ * the GNU General Public License Version 2. See the LICENSE file
+ * at the top of the source tree.
+ */
+
+/*!
+ * \file
+ * \brief Opus format attribute interface
+ *
+ * \author Lorenzo Miniero <lorenzo@meetecho.com>
+ */
+
+/*** MODULEINFO
+	<support_level>core</support_level>
+ ***/
+
+#include "asterisk.h"
+
+ASTERISK_FILE_VERSION(__FILE__, "$Revision$")
+
+#include "asterisk/module.h"
+#include "asterisk/format.h"
+
+/*!
+ * \brief Opus attribute structure.
+ *
+ * \note http://tools.ietf.org/html/draft-ietf-payload-rtp-opus-00.
+ */
+struct opus_attr {
+	unsigned int maxbitrate;	        /* Default 64-128 kb/s for FB stereo music */
+	unsigned int maxplayrate	        /* Default 48000 */;
+	unsigned int minptime;	            /* Default 3, but it's 10 in format.c */
+	unsigned int stereo;	            /* Default 0 */
+	unsigned int cbr;	                /* Default 0 */
+	unsigned int fec;	                /* Default 0 */
+	unsigned int dtx;	                /* Default 0 */
+	unsigned int spropmaxcapturerate;	/* Default 48000 */
+	unsigned int spropstereo;	        /* Default 0 */
+};
+
+static int opus_sdp_parse(struct ast_format_attr *format_attr, const char *attributes)
+{
+	struct opus_attr *attr = (struct opus_attr *) format_attr;
+	const char *kvp;
+	unsigned int val;
+
+	if ((kvp = strstr(attributes, "maxplaybackrate")) && sscanf(kvp, "maxplaybackrate=%30u", &val) == 1) {
+		attr->maxplayrate = val;
+	}
+	if ((kvp = strstr(attributes, "sprop-maxcapturerate")) && sscanf(kvp, "sprop-maxcapturerate=%30u", &val) == 1) {
+		attr->spropmaxcapturerate = val;
+	}
+	if ((kvp = strstr(attributes, "minptime")) && sscanf(kvp, "minptime=%30u", &val) == 1) {
+		attr->minptime = val;
+	}
+	if ((kvp = strstr(attributes, "maxaveragebitrate")) && sscanf(kvp, "maxaveragebitrate=%30u", &val) == 1) {
+		attr->maxbitrate = val;
+	}
+	if ((kvp = strstr(attributes, " stereo")) && sscanf(kvp, " stereo=%30u", &val) == 1) {
+		attr->stereo = val;
+	}
+	if ((kvp = strstr(attributes, ";stereo")) && sscanf(kvp, ";stereo=%30u", &val) == 1) {
+		attr->stereo = val;
+	}
+	if ((kvp = strstr(attributes, "sprop-stereo")) && sscanf(kvp, "sprop-stereo=%30u", &val) == 1) {
+		attr->spropstereo = val;
+	}
+	if ((kvp = strstr(attributes, "cbr")) && sscanf(kvp, "cbr=%30u", &val) == 1) {
+		attr->cbr = val;
+	}
+	if ((kvp = strstr(attributes, "useinbandfec")) && sscanf(kvp, "useinbandfec=%30u", &val) == 1) {
+		attr->fec = val;
+	}
+	if ((kvp = strstr(attributes, "usedtx")) && sscanf(kvp, "usedtx=%30u", &val) == 1) {
+		attr->dtx = val;
+	}
+
+	return 0;
+}
+
+static void opus_sdp_generate(const struct ast_format_attr *format_attr, unsigned int payload, struct ast_str **str)
+{
+	struct opus_attr *attr = (struct opus_attr *) format_attr;
+
+	/* FIXME should we only generate attributes that were explicitly set? */
+	ast_str_append(str, 0,
+				"a=fmtp:%d "
+					"maxplaybackrate=%d;"
+					"sprop-maxcapturerate=%d;"
+					"minptime=%d;"
+					"maxaveragebitrate=%d;"
+					"stereo=%d;"
+					"sprop-stereo=%d;"
+					"cbr=%d;"
+					"useinbandfec=%d;"
+					"usedtx=%d\r\n",
+			payload,
+			attr->maxplayrate ? attr->maxplayrate : 48000,	/* maxplaybackrate */
+			attr->spropmaxcapturerate ? attr->spropmaxcapturerate : 48000,	/* sprop-maxcapturerate */
+			attr->minptime > 10 ? attr->minptime : 10,	/* minptime */
+			attr->maxbitrate ? attr->maxbitrate : 20000,	/* maxaveragebitrate */
+			attr->stereo ? 1 : 0,		/* stereo */
+			attr->spropstereo ? 1 : 0,		/* sprop-stereo */
+			attr->cbr ? 1 : 0,		/* cbr */
+			attr->fec ? 1 : 0,		/* useinbandfec */
+			attr->dtx ? 1 : 0		/* usedtx */
+	);
+}
+
+static int opus_get_val(const struct ast_format_attr *fattr, int key, void *result)
+{
+	const struct opus_attr *attr = (struct opus_attr *) fattr;
+	int *val = result;
+
+	switch (key) {
+	case OPUS_ATTR_KEY_MAX_BITRATE:
+		*val = attr->maxbitrate;
+		break;
+	case OPUS_ATTR_KEY_MAX_PLAYRATE:
+		*val = attr->maxplayrate;
+		break;
+	case OPUS_ATTR_KEY_MINPTIME:
+		*val = attr->minptime;
+		break;
+	case OPUS_ATTR_KEY_STEREO:
+		*val = attr->stereo;
+		break;
+	case OPUS_ATTR_KEY_CBR:
+		*val = attr->cbr;
+		break;
+	case OPUS_ATTR_KEY_FEC:
+		*val = attr->fec;
+		break;
+	case OPUS_ATTR_KEY_DTX:
+		*val = attr->dtx;
+		break;
+	case OPUS_ATTR_KEY_SPROP_CAPTURE_RATE:
+		*val = attr->spropmaxcapturerate;
+		break;
+	case OPUS_ATTR_KEY_SPROP_STEREO:
+		*val = attr->spropstereo;
+		break;
+	default:
+		ast_log(LOG_WARNING, "unknown attribute type %d\n", key);
+		return -1;
+	}
+	return 0;
+}
+
+static int opus_isset(const struct ast_format_attr *fattr, va_list ap)
+{
+	enum opus_attr_keys key;
+	const struct opus_attr *attr = (struct opus_attr *) fattr;
+
+	for (key = va_arg(ap, int);
+		key != AST_FORMAT_ATTR_END;
+		key = va_arg(ap, int))
+	{
+		switch (key) {
+		case OPUS_ATTR_KEY_MAX_BITRATE:
+			if (attr->maxbitrate != (va_arg(ap, int))) {
+				return -1;
+			}
+			break;
+		case OPUS_ATTR_KEY_MAX_PLAYRATE:
+			if (attr->maxplayrate != (va_arg(ap, int))) {
+				return -1;
+			}
+			break;
+		case OPUS_ATTR_KEY_MINPTIME:
+			if (attr->minptime != (va_arg(ap, int))) {
+				return -1;
+			}
+			break;
+		case OPUS_ATTR_KEY_STEREO:
+			if (attr->stereo != (va_arg(ap, int))) {
+				return -1;
+			}
+			break;
+		case OPUS_ATTR_KEY_CBR:
+			if (attr->cbr != (va_arg(ap, int))) {
+				return -1;
+			}
+			break;
+		case OPUS_ATTR_KEY_FEC:
+			if (attr->fec != (va_arg(ap, int))) {
+				return -1;
+			}
+			break;
+		case OPUS_ATTR_KEY_DTX:
+			if (attr->dtx != (va_arg(ap, int))) {
+				return -1;
+			}
+			break;
+		case OPUS_ATTR_KEY_SPROP_CAPTURE_RATE:
+			if (attr->spropmaxcapturerate != (va_arg(ap, int))) {
+				return -1;
+			}
+			break;
+		case OPUS_ATTR_KEY_SPROP_STEREO:
+			if (attr->spropstereo != (va_arg(ap, int))) {
+				return -1;
+			}
+			break;
+		default:
+			ast_log(LOG_WARNING, "unknown attribute type %d\n", key);
+			return -1;
+		}
+	}
+	return 0;
+}
+static int opus_getjoint(const struct ast_format_attr *fattr1, const struct ast_format_attr *fattr2, struct ast_format_attr *result)
+{
+	struct opus_attr *attr1 = (struct opus_attr *) fattr1;
+	struct opus_attr *attr2 = (struct opus_attr *) fattr2;
+	struct opus_attr *attr_res = (struct opus_attr *) result;
+	int joint = 0;
+
+	/* Only do dtx if both sides want it. DTX is a trade off between
+	 * computational complexity and bandwidth. */
+	attr_res->dtx = attr1->dtx && attr2->dtx ? 1 : 0;
+
+	/* Only do FEC if both sides want it.  If a peer specifically requests not
+	 * to receive with FEC, it may be a waste of bandwidth. */
+	attr_res->fec = attr1->fec && attr2->fec ? 1 : 0;
+
+	/* Only do stereo if both sides want it.  If a peer specifically requests not
+	 * to receive stereo signals, it may be a waste of bandwidth. */
+	attr_res->stereo = attr1->stereo && attr2->stereo ? 1 : 0;
+
+	/* FIXME: do we need to join other attributes as well, e.g., minptime, cbr, etc.? */
+
+	return joint;
+}
+
+static void opus_set(struct ast_format_attr *fattr, va_list ap)
+{
+	enum opus_attr_keys key;
+	struct opus_attr *attr = (struct opus_attr *) fattr;
+
+	for (key = va_arg(ap, int);
+		key != AST_FORMAT_ATTR_END;
+		key = va_arg(ap, int))
+	{
+		switch (key) {
+		case OPUS_ATTR_KEY_MAX_BITRATE:
+			attr->maxbitrate = (va_arg(ap, int));
+			break;
+		case OPUS_ATTR_KEY_MAX_PLAYRATE:
+			attr->maxplayrate = (va_arg(ap, int));
+			break;
+		case OPUS_ATTR_KEY_MINPTIME:
+			attr->minptime = (va_arg(ap, int));
+			break;
+		case OPUS_ATTR_KEY_STEREO:
+			attr->stereo = (va_arg(ap, int));
+			break;
+		case OPUS_ATTR_KEY_CBR:
+			attr->cbr = (va_arg(ap, int));
+			break;
+		case OPUS_ATTR_KEY_FEC:
+			attr->fec = (va_arg(ap, int));
+			break;
+		case OPUS_ATTR_KEY_DTX:
+			attr->dtx = (va_arg(ap, int));
+			break;
+		case OPUS_ATTR_KEY_SPROP_CAPTURE_RATE:
+			attr->spropmaxcapturerate = (va_arg(ap, int));
+			break;
+		case OPUS_ATTR_KEY_SPROP_STEREO:
+			attr->spropstereo = (va_arg(ap, int));
+			break;
+		default:
+			ast_log(LOG_WARNING, "unknown attribute type %d\n", key);
+		}
+	}
+}
+
+static struct ast_format_attr_interface opus_interface = {
+	.id = AST_FORMAT_OPUS,
+	.format_attr_get_joint = opus_getjoint,
+	.format_attr_set = opus_set,
+	.format_attr_isset = opus_isset,
+	.format_attr_get_val = opus_get_val,
+	.format_attr_sdp_parse = opus_sdp_parse,
+	.format_attr_sdp_generate = opus_sdp_generate,
+};
+
+static int load_module(void)
+{
+	if (ast_format_attr_reg_interface(&opus_interface)) {
+		return AST_MODULE_LOAD_DECLINE;
+	}
+
+	return AST_MODULE_LOAD_SUCCESS;
+}
+
+static int unload_module(void)
+{
+	ast_format_attr_unreg_interface(&opus_interface);
+	return 0;
+}
+
+AST_MODULE_INFO(ASTERISK_GPL_KEY, AST_MODFLAG_LOAD_ORDER, "Opus Format Attribute Module",
+	.load = load_module,
+	.unload = unload_module,
+	.load_pri = AST_MODPRI_CHANNEL_DEPEND,
+);
diff --git a/res/res_pjsip_sdp_rtp.c b/res/res_pjsip_sdp_rtp.c
index c97c0cb40..be5d59f06 100644
--- a/res/res_pjsip_sdp_rtp.c
+++ b/res/res_pjsip_sdp_rtp.c
@@ -849,7 +849,9 @@ static int create_outgoing_sdp_stream(struct ast_sip_session *session, struct as
 	char tmp[512];
 	pj_str_t stmp;
 	pjmedia_sdp_attr *attr;
-	int index = 0, min_packet_size = 0, noncodec = (session->endpoint->dtmf == AST_SIP_DTMF_RFC_4733) ? AST_RTP_DTMF : 0;
+	int index = 0;
+	int noncodec = (session->endpoint->dtmf == AST_SIP_DTMF_RFC_4733) ? AST_RTP_DTMF : 0;
+	int min_packet_size = 0, max_packet_size = 0;
 	int rtp_code;
 	struct ast_format format;
 	RAII_VAR(struct ast_format_cap *, caps, NULL, ast_format_cap_destroy);
@@ -951,6 +953,10 @@ static int create_outgoing_sdp_stream(struct ast_sip_session *session, struct as
 			if (fmt.cur_ms && ((fmt.cur_ms < min_packet_size) || !min_packet_size)) {
 				min_packet_size = fmt.cur_ms;
 			}
+
+			if (fmt.max_ms && ((fmt.max_ms < max_packet_size) || !max_packet_size)) {
+				max_packet_size = fmt.max_ms;
+			}
 		}
 	}
 
@@ -983,6 +989,12 @@ static int create_outgoing_sdp_stream(struct ast_sip_session *session, struct as
 		media->attr[media->attr_count++] = attr;
 	}
 
+	if (max_packet_size) {
+		snprintf(tmp, sizeof(tmp), "%d", max_packet_size);
+		attr = pjmedia_sdp_attr_create(pool, "maxptime", pj_cstr(&stmp, tmp));
+		media->attr[media->attr_count++] = attr;
+	}
+
 	/* Add the sendrecv attribute - we purposely don't keep track because pjmedia-sdp will automatically change our offer for us */
 	attr = PJ_POOL_ZALLOC_T(pool, pjmedia_sdp_attr);
 	attr->name = STR_SENDRECV;
diff --git a/res/res_rtp_asterisk.c b/res/res_rtp_asterisk.c
index db07e4ec5..6383b09e3 100644
--- a/res/res_rtp_asterisk.c
+++ b/res/res_rtp_asterisk.c
@@ -90,6 +90,8 @@ ASTERISK_FILE_VERSION(__FILE__, "$Revision$")
 #define RTCP_PT_SDES    202
 #define RTCP_PT_BYE     203
 #define RTCP_PT_APP     204
+/* VP8: RTCP Feedback */
+#define RTCP_PT_PSFB    206
 
 #define RTP_MTU		1200
 
@@ -350,6 +352,9 @@ struct ast_rtcp {
 	double normdevrtt;
 	double stdevrtt;
 	unsigned int rtt_count;
+
+	/* VP8: sequence number for the RTCP FIR FCI */
+	int firseq;
 };
 
 struct rtp_red {
@@ -2414,7 +2419,7 @@ static int ast_rtcp_write(const void *data)
 	}
 
 	if (!res) {
-		/* 
+		/*
 		 * Not being rescheduled.
 		 */
 		ao2_ref(instance, -1);
@@ -2609,6 +2614,45 @@ static int ast_rtp_write(struct ast_rtp_instance *instance, struct ast_frame *fr
 		return 0;
 	}
 
+	/* VP8: is this a request to send a RTCP FIR? */
+	if (frame->frametype == AST_FRAME_CONTROL && frame->subclass.integer == AST_CONTROL_VIDUPDATE) {
+		struct ast_rtp *rtp = ast_rtp_instance_get_data(instance);
+		unsigned int *rtcpheader;
+		char bdata[1024];
+		int len = 20;
+		int ice;
+		int res;
+
+		if (!rtp || !rtp->rtcp) {
+			return 0;
+		}
+
+		if (ast_sockaddr_isnull(&rtp->rtcp->them)) {
+			/*
+			 * RTCP was stopped.
+			 */
+			return 0;
+		}
+
+		/* Prepare RTCP FIR (PT=206, FMT=4) */
+		rtp->rtcp->firseq++;
+		if(rtp->rtcp->firseq == 256) {
+			rtp->rtcp->firseq = 0;
+		}
+
+		rtcpheader = (unsigned int *)bdata;
+		rtcpheader[0] = htonl((2 << 30) | (4 << 24) | (RTCP_PT_PSFB << 16) | ((len/4)-1));
+		rtcpheader[1] = htonl(rtp->ssrc);
+		rtcpheader[2] = htonl(rtp->themssrc);
+		rtcpheader[3] = htonl(rtp->themssrc);	/* FCI: SSRC */
+		rtcpheader[4] = htonl(rtp->rtcp->firseq << 24);			/* FCI: Sequence number */
+		res = rtcp_sendto(instance, (unsigned int *)rtcpheader, len, 0, &rtp->rtcp->them, &ice);
+		if (res < 0) {
+			ast_log(LOG_ERROR, "RTCP FIR transmission error: %s\n", strerror(errno));
+		}
+		return 0;
+	}
+
 	/* If there is no data length we can't very well send the packet */
 	if (!frame->datalen) {
 		ast_debug(1, "Received frame with no data for RTP instance '%p' so dropping frame\n", instance);
@@ -2660,6 +2704,8 @@ static int ast_rtp_write(struct ast_rtp_instance *instance, struct ast_frame *fr
 		case AST_FORMAT_SIREN7:
 		case AST_FORMAT_SIREN14:
 		case AST_FORMAT_G719:
+		/* Opus */
+		case AST_FORMAT_OPUS:
 			/* these are all frame-based codecs and cannot be safely run through
 			   a smoother */
 			break;
@@ -3353,6 +3399,8 @@ static struct ast_frame *ast_rtcp_read(struct ast_rtp_instance *instance)
 					message_blob);
 			break;
 		case RTCP_PT_FUR:
+		/* Handle RTCP FIR as FUR */
+		case RTCP_PT_PSFB:
 			if (rtcp_debug_test_addr(&addr)) {
 				ast_verbose("Received an RTCP Fast Update Request\n");
 			}
@@ -4174,14 +4222,14 @@ static int ast_rtp_sendcng(struct ast_rtp_instance *instance, int level)
 	payload = ast_rtp_codecs_payload_code(ast_rtp_instance_get_codecs(instance), 0, NULL, AST_RTP_CN);
 
 	level = 127 - (level & 0x7f);
-	
+
 	rtp->dtmfmute = ast_tvadd(ast_tvnow(), ast_tv(0, 500000));
 
 	/* Get a pointer to the header */
 	rtpheader = (unsigned int *)data;
 	rtpheader[0] = htonl((2 << 30) | (payload << 16) | (rtp->seqno));
 	rtpheader[1] = htonl(rtp->lastts);
-	rtpheader[2] = htonl(rtp->ssrc); 
+	rtpheader[2] = htonl(rtp->ssrc);
 	data[12] = level;
 
 	res = rtp_sendto(instance, (void *) rtpheader, hdrlen + 1, 0, &remote_address, &ice);