create libosmo-mgcp and osmo-mgw by copying legacy code

This a cosmetic commit, copying libosmo-legacy-mgcp to libosmo-mgcp and
osmo-bsc_mgcp to osmo-mgw 1:1 at first, to provide a basis for next patches
that highlight the changes from legacy to new code.

Until osmo-msc and osmo-bsc are adjusted to operate with the new code, we will
keep the legacy code alongside the new code. The legacy code might be dropped
later.

Change-Id: Idf54481754a1765bdb2d0d7033bc0d7dc2018024
diff --git a/src/libosmo-mgcp/mgcp_network.c b/src/libosmo-mgcp/mgcp_network.c
new file mode 100644
index 0000000..8ccfb42
--- /dev/null
+++ b/src/libosmo-mgcp/mgcp_network.c
@@ -0,0 +1,1166 @@
+/* A Media Gateway Control Protocol Media Gateway: RFC 3435 */
+/* The protocol implementation */
+
+/*
+ * (C) 2009-2012 by Holger Hans Peter Freyther <zecke@selfish.org>
+ * (C) 2009-2012 by On-Waves
+ * All Rights Reserved
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU Affero General Public License as published by
+ * the Free Software Foundation; either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * This program 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 Affero General Public License for more details.
+ *
+ * You should have received a copy of the GNU Affero General Public License
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
+ *
+ */
+
+#include <string.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <errno.h>
+#include <time.h>
+#include <limits.h>
+
+#include <sys/socket.h>
+#include <arpa/inet.h>
+
+#include <osmocom/core/msgb.h>
+#include <osmocom/core/select.h>
+
+#include <osmocom/netif/rtp.h>
+
+#include <osmocom/legacy_mgcp/mgcp.h>
+#include <osmocom/legacy_mgcp/mgcp_internal.h>
+
+#include <osmocom/legacy_mgcp/osmux.h>
+
+#warning "Make use of the rtp proxy code"
+
+
+#define RTP_SEQ_MOD		(1 << 16)
+#define RTP_MAX_DROPOUT		3000
+#define RTP_MAX_MISORDER	100
+#define RTP_BUF_SIZE		4096
+
+enum {
+	MGCP_PROTO_RTP,
+	MGCP_PROTO_RTCP,
+};
+
+/**
+ * This does not need to be a precision timestamp and
+ * is allowed to wrap quite fast. The returned value is
+ * 1/unit seconds.
+ */
+static uint32_t get_current_ts(unsigned unit)
+{
+	struct timespec tp;
+	uint64_t ret;
+
+	if (!unit)
+		return 0;
+
+	memset(&tp, 0, sizeof(tp));
+	if (clock_gettime(CLOCK_MONOTONIC, &tp) != 0)
+		LOGP(DLMGCP, LOGL_NOTICE,
+			"Getting the clock failed.\n");
+
+	/* convert it to 1/unit seconds */
+	ret = tp.tv_sec;
+	ret *= unit;
+	ret += (int64_t)tp.tv_nsec * unit / 1000 / 1000 / 1000;
+
+	return ret;
+}
+
+int mgcp_udp_send(int fd, struct in_addr *addr, int port, char *buf, int len)
+{
+	struct sockaddr_in out;
+	out.sin_family = AF_INET;
+	out.sin_port = port;
+	memcpy(&out.sin_addr, addr, sizeof(*addr));
+
+	return sendto(fd, buf, len, 0, (struct sockaddr *)&out, sizeof(out));
+}
+
+int mgcp_send_dummy(struct mgcp_endpoint *endp)
+{
+	static char buf[] = { MGCP_DUMMY_LOAD };
+	int rc;
+	int was_rtcp = 0;
+
+	rc = mgcp_udp_send(endp->net_end.rtp.fd, &endp->net_end.addr,
+			   endp->net_end.rtp_port, buf, 1);
+
+	if (rc == -1)
+		goto failed;
+
+	if (endp->tcfg->omit_rtcp)
+		return rc;
+
+	was_rtcp = 1;
+	rc = mgcp_udp_send(endp->net_end.rtcp.fd, &endp->net_end.addr,
+			   endp->net_end.rtcp_port, buf, 1);
+
+	if (rc >= 0)
+		return rc;
+
+failed:
+	LOGP(DLMGCP, LOGL_ERROR,
+		"Failed to send dummy %s packet: %s on: 0x%x to %s:%d\n",
+		was_rtcp ? "RTCP" : "RTP",
+		strerror(errno), ENDPOINT_NUMBER(endp), inet_ntoa(endp->net_end.addr),
+		was_rtcp ? endp->net_end.rtcp_port : endp->net_end.rtp_port);
+
+	return -1;
+}
+
+static int32_t compute_timestamp_aligment_error(struct mgcp_rtp_stream_state *sstate,
+						int ptime, uint32_t timestamp)
+{
+	int32_t timestamp_delta;
+
+	if (ptime == 0)
+		return 0;
+
+	/* Align according to: T - Tlast = k * Tptime */
+	timestamp_delta = timestamp - sstate->last_timestamp;
+
+	return timestamp_delta % ptime;
+}
+
+static int check_rtp_timestamp(struct mgcp_endpoint *endp,
+			       struct mgcp_rtp_state *state,
+			       struct mgcp_rtp_stream_state *sstate,
+			       struct mgcp_rtp_end *rtp_end,
+			       struct sockaddr_in *addr,
+			       uint16_t seq, uint32_t timestamp,
+			       const char *text, int32_t *tsdelta_out)
+{
+	int32_t tsdelta;
+	int32_t timestamp_error;
+
+	/* Not fully intialized, skip */
+	if (sstate->last_tsdelta == 0 && timestamp == sstate->last_timestamp)
+		return 0;
+
+	if (seq == sstate->last_seq) {
+		if (timestamp != sstate->last_timestamp) {
+			sstate->err_ts_counter += 1;
+			LOGP(DLMGCP, LOGL_ERROR,
+			     "The %s timestamp delta is != 0 but the sequence "
+			     "number %d is the same, "
+			     "TS offset: %d, SeqNo offset: %d "
+			     "on 0x%x SSRC: %u timestamp: %u "
+			     "from %s:%d in %d\n",
+			     text, seq,
+			     state->timestamp_offset, state->seq_offset,
+			     ENDPOINT_NUMBER(endp), sstate->ssrc, timestamp,
+			     inet_ntoa(addr->sin_addr), ntohs(addr->sin_port),
+			     endp->conn_mode);
+		}
+		return 0;
+	}
+
+	tsdelta =
+		(int32_t)(timestamp - sstate->last_timestamp) /
+		(int16_t)(seq - sstate->last_seq);
+
+	if (tsdelta == 0) {
+		/* Don't update *tsdelta_out */
+		LOGP(DLMGCP, LOGL_NOTICE,
+		     "The %s timestamp delta is %d "
+		     "on 0x%x SSRC: %u timestamp: %u "
+		     "from %s:%d in %d\n",
+		     text, tsdelta,
+		     ENDPOINT_NUMBER(endp), sstate->ssrc, timestamp,
+		     inet_ntoa(addr->sin_addr), ntohs(addr->sin_port),
+		     endp->conn_mode);
+
+		return 0;
+	}
+
+	if (sstate->last_tsdelta != tsdelta) {
+		if (sstate->last_tsdelta) {
+			LOGP(DLMGCP, LOGL_INFO,
+			     "The %s timestamp delta changes from %d to %d "
+			     "on 0x%x SSRC: %u timestamp: %u from %s:%d in %d\n",
+			     text, sstate->last_tsdelta, tsdelta,
+			     ENDPOINT_NUMBER(endp), sstate->ssrc, timestamp,
+			     inet_ntoa(addr->sin_addr), ntohs(addr->sin_port),
+			     endp->conn_mode);
+		}
+	}
+
+	if (tsdelta_out)
+		*tsdelta_out = tsdelta;
+
+	timestamp_error =
+		compute_timestamp_aligment_error(sstate, state->packet_duration,
+						 timestamp);
+
+	if (timestamp_error) {
+		sstate->err_ts_counter += 1;
+		LOGP(DLMGCP, LOGL_NOTICE,
+		     "The %s timestamp has an alignment error of %d "
+		     "on 0x%x SSRC: %u "
+		     "SeqNo delta: %d, TS delta: %d, dTS/dSeq: %d "
+		     "from %s:%d in mode %d. ptime: %d\n",
+		     text, timestamp_error,
+		     ENDPOINT_NUMBER(endp), sstate->ssrc,
+		     (int16_t)(seq - sstate->last_seq),
+		     (int32_t)(timestamp - sstate->last_timestamp),
+		     tsdelta,
+		     inet_ntoa(addr->sin_addr), ntohs(addr->sin_port),
+		     endp->conn_mode, state->packet_duration);
+	}
+	return 1;
+}
+
+/* Set the timestamp offset according to the packet duration. */
+static int adjust_rtp_timestamp_offset(struct mgcp_endpoint *endp,
+				       struct mgcp_rtp_state *state,
+				       struct mgcp_rtp_end *rtp_end,
+				       struct sockaddr_in *addr,
+				       int16_t delta_seq, uint32_t in_timestamp)
+{
+	int32_t tsdelta = state->packet_duration;
+	int timestamp_offset;
+	uint32_t out_timestamp;
+
+	if (tsdelta == 0) {
+		tsdelta = state->out_stream.last_tsdelta;
+		if (tsdelta != 0) {
+			LOGP(DLMGCP, LOGL_NOTICE,
+			     "A fixed packet duration is not available on 0x%x, "
+			     "using last output timestamp delta instead: %d "
+			     "from %s:%d in %d\n",
+			     ENDPOINT_NUMBER(endp), tsdelta,
+			     inet_ntoa(addr->sin_addr), ntohs(addr->sin_port),
+			     endp->conn_mode);
+		} else {
+			tsdelta = rtp_end->codec.rate * 20 / 1000;
+			LOGP(DLMGCP, LOGL_NOTICE,
+			     "Fixed packet duration and last timestamp delta "
+			     "are not available on 0x%x, "
+			     "using fixed 20ms instead: %d "
+			     "from %s:%d in %d\n",
+			     ENDPOINT_NUMBER(endp), tsdelta,
+			     inet_ntoa(addr->sin_addr), ntohs(addr->sin_port),
+			     endp->conn_mode);
+		}
+	}
+
+	out_timestamp = state->out_stream.last_timestamp + delta_seq * tsdelta;
+	timestamp_offset = out_timestamp - in_timestamp;
+
+	if (state->timestamp_offset != timestamp_offset) {
+		state->timestamp_offset = timestamp_offset;
+
+		LOGP(DLMGCP, LOGL_NOTICE,
+		     "Timestamp offset change on 0x%x SSRC: %u "
+		     "SeqNo delta: %d, TS offset: %d, "
+		     "from %s:%d in %d\n",
+		     ENDPOINT_NUMBER(endp), state->in_stream.ssrc,
+		     delta_seq, state->timestamp_offset,
+		     inet_ntoa(addr->sin_addr), ntohs(addr->sin_port),
+		     endp->conn_mode);
+	}
+
+	return timestamp_offset;
+}
+
+/* Set the timestamp offset according to the packet duration. */
+static int align_rtp_timestamp_offset(struct mgcp_endpoint *endp,
+				      struct mgcp_rtp_state *state,
+				      struct mgcp_rtp_end *rtp_end,
+				      struct sockaddr_in *addr,
+				      uint32_t timestamp)
+{
+	int timestamp_error = 0;
+	int ptime = state->packet_duration;
+
+	/* Align according to: T + Toffs - Tlast = k * Tptime */
+
+	timestamp_error = compute_timestamp_aligment_error(
+		&state->out_stream, ptime,
+		timestamp + state->timestamp_offset);
+
+	if (timestamp_error) {
+		state->timestamp_offset += ptime - timestamp_error;
+
+		LOGP(DLMGCP, LOGL_NOTICE,
+		     "Corrected timestamp alignment error of %d on 0x%x SSRC: %u "
+		     "new TS offset: %d, "
+		     "from %s:%d in %d\n",
+		     timestamp_error,
+		     ENDPOINT_NUMBER(endp), state->in_stream.ssrc,
+		     state->timestamp_offset, inet_ntoa(addr->sin_addr),
+		     ntohs(addr->sin_port), endp->conn_mode);
+	}
+
+	OSMO_ASSERT(compute_timestamp_aligment_error(&state->out_stream, ptime,
+						     timestamp + state->timestamp_offset) == 0);
+
+	return timestamp_error;
+}
+
+int mgcp_rtp_processing_default(struct mgcp_endpoint *endp, struct mgcp_rtp_end *dst_end,
+				char *data, int *len, int buf_size)
+{
+	return 0;
+}
+
+int mgcp_setup_rtp_processing_default(struct mgcp_endpoint *endp,
+				      struct mgcp_rtp_end *dst_end,
+				      struct mgcp_rtp_end *src_end)
+{
+	return 0;
+}
+
+void mgcp_get_net_downlink_format_default(struct mgcp_endpoint *endp,
+					  int *payload_type,
+					  const char**audio_name,
+					  const char**fmtp_extra)
+{
+	/* Use the BTS side parameters when passing the SDP data (for
+	 * downlink) to the net peer.
+	 */
+	*payload_type = endp->bts_end.codec.payload_type;
+	*audio_name = endp->bts_end.codec.audio_name;
+	*fmtp_extra = endp->bts_end.fmtp_extra;
+}
+
+
+void mgcp_rtp_annex_count(struct mgcp_endpoint *endp, struct mgcp_rtp_state *state,
+			const uint16_t seq, const int32_t transit,
+			const uint32_t ssrc)
+{
+	int32_t d;
+
+	/* initialize or re-initialize */
+	if (!state->stats_initialized || state->stats_ssrc != ssrc) {
+		state->stats_initialized = 1;
+		state->stats_base_seq = seq;
+		state->stats_max_seq = seq - 1;
+		state->stats_ssrc = ssrc;
+		state->stats_jitter = 0;
+		state->stats_transit = transit;
+		state->stats_cycles = 0;
+	} else {
+		uint16_t udelta;
+
+		/*
+		 * The below takes the shape of the validation of
+		 * Appendix A. Check if there is something weird with
+		 * the sequence number, otherwise check for a wrap
+		 * around in the sequence number.
+		 * It can't wrap during the initialization so let's
+		 * skip it here. The Appendix A probably doesn't have
+		 * this issue because of the probation.
+		 */
+		udelta = seq - state->stats_max_seq;
+		if (udelta < RTP_MAX_DROPOUT) {
+			if (seq < state->stats_max_seq)
+				state->stats_cycles += RTP_SEQ_MOD;
+		} else if (udelta <= RTP_SEQ_MOD - RTP_MAX_MISORDER) {
+			LOGP(DLMGCP, LOGL_NOTICE,
+				"RTP seqno made a very large jump on 0x%x delta: %u\n",
+				ENDPOINT_NUMBER(endp), udelta);
+		}
+	}
+
+	/*
+	 * Calculate the jitter between the two packages. The TS should be
+	 * taken closer to the read function. This was taken from the
+	 * Appendix A of RFC 3550. Timestamp and arrival_time have a 1/rate
+	 * resolution.
+	 */
+	d = transit - state->stats_transit;
+	state->stats_transit = transit;
+	if (d < 0)
+		d = -d;
+	state->stats_jitter += d - ((state->stats_jitter + 8) >> 4);
+	state->stats_max_seq = seq;
+}
+
+
+
+/**
+ * The RFC 3550 Appendix A assumes there are multiple sources but
+ * some of the supported endpoints (e.g. the nanoBTS) can only handle
+ * one source and this code will patch RTP header to appear as if there
+ * is only one source.
+ * There is also no probation period for new sources. Every RTP header
+ * we receive will be seen as a switch in streams.
+ */
+void mgcp_patch_and_count(struct mgcp_endpoint *endp, struct mgcp_rtp_state *state,
+			  struct mgcp_rtp_end *rtp_end, struct sockaddr_in *addr,
+			  char *data, int len)
+{
+	uint32_t arrival_time;
+	int32_t transit;
+	uint16_t seq;
+	uint32_t timestamp, ssrc;
+	struct rtp_hdr *rtp_hdr;
+	int payload = rtp_end->codec.payload_type;
+
+	if (len < sizeof(*rtp_hdr))
+		return;
+
+	rtp_hdr = (struct rtp_hdr *) data;
+	seq = ntohs(rtp_hdr->sequence);
+	timestamp = ntohl(rtp_hdr->timestamp);
+	arrival_time = get_current_ts(rtp_end->codec.rate);
+	ssrc = ntohl(rtp_hdr->ssrc);
+	transit = arrival_time - timestamp;
+
+	mgcp_rtp_annex_count(endp, state, seq, transit, ssrc);
+
+	if (!state->initialized) {
+		state->initialized = 1;
+		state->in_stream.last_seq = seq - 1;
+		state->in_stream.ssrc = state->orig_ssrc = ssrc;
+		state->in_stream.last_tsdelta = 0;
+		state->packet_duration = mgcp_rtp_packet_duration(endp, rtp_end);
+		state->out_stream = state->in_stream;
+		state->out_stream.last_timestamp = timestamp;
+		state->out_stream.ssrc = ssrc - 1; /* force output SSRC change */
+		LOGP(DLMGCP, LOGL_INFO,
+			"Initializing stream on 0x%x SSRC: %u timestamp: %u "
+			"pkt-duration: %d, from %s:%d in %d\n",
+			ENDPOINT_NUMBER(endp), state->in_stream.ssrc,
+			state->seq_offset, state->packet_duration,
+			inet_ntoa(addr->sin_addr), ntohs(addr->sin_port),
+			endp->conn_mode);
+		if (state->packet_duration == 0) {
+			state->packet_duration = rtp_end->codec.rate * 20 / 1000;
+			LOGP(DLMGCP, LOGL_NOTICE,
+			     "Fixed packet duration is not available on 0x%x, "
+			     "using fixed 20ms instead: %d from %s:%d in %d\n",
+			     ENDPOINT_NUMBER(endp), state->packet_duration,
+			     inet_ntoa(addr->sin_addr), ntohs(addr->sin_port),
+			     endp->conn_mode);
+		}
+	} else if (state->in_stream.ssrc != ssrc) {
+		LOGP(DLMGCP, LOGL_NOTICE,
+			"The SSRC changed on 0x%x: %u -> %u  "
+			"from %s:%d in %d\n",
+			ENDPOINT_NUMBER(endp),
+			state->in_stream.ssrc, rtp_hdr->ssrc,
+			inet_ntoa(addr->sin_addr), ntohs(addr->sin_port),
+			endp->conn_mode);
+
+		state->in_stream.ssrc = ssrc;
+		if (rtp_end->force_constant_ssrc) {
+			int16_t delta_seq;
+
+			/* Always increment seqno by 1 */
+			state->seq_offset =
+				(state->out_stream.last_seq + 1) - seq;
+
+			/* Estimate number of packets that would have been sent */
+			delta_seq =
+				(arrival_time - state->in_stream.last_arrival_time
+				 + state->packet_duration/2) /
+				state->packet_duration;
+
+			adjust_rtp_timestamp_offset(endp, state, rtp_end, addr,
+						    delta_seq, timestamp);
+
+			state->patch_ssrc = 1;
+			ssrc = state->orig_ssrc;
+			if (rtp_end->force_constant_ssrc != -1)
+				rtp_end->force_constant_ssrc -= 1;
+
+			LOGP(DLMGCP, LOGL_NOTICE,
+			     "SSRC patching enabled on 0x%x SSRC: %u "
+			     "SeqNo offset: %d, TS offset: %d "
+			     "from %s:%d in %d\n",
+			     ENDPOINT_NUMBER(endp), state->in_stream.ssrc,
+			     state->seq_offset, state->timestamp_offset,
+			     inet_ntoa(addr->sin_addr), ntohs(addr->sin_port),
+			     endp->conn_mode);
+		}
+
+		state->in_stream.last_tsdelta = 0;
+	} else {
+		/* Compute current per-packet timestamp delta */
+		check_rtp_timestamp(endp, state, &state->in_stream, rtp_end, addr,
+				    seq, timestamp, "input",
+				    &state->in_stream.last_tsdelta);
+
+		if (state->patch_ssrc)
+			ssrc = state->orig_ssrc;
+	}
+
+	/* Save before patching */
+	state->in_stream.last_timestamp = timestamp;
+	state->in_stream.last_seq = seq;
+	state->in_stream.last_arrival_time = arrival_time;
+
+	if (rtp_end->force_aligned_timing &&
+	    state->out_stream.ssrc == ssrc && state->packet_duration)
+		/* Align the timestamp offset */
+		align_rtp_timestamp_offset(endp, state, rtp_end, addr, timestamp);
+
+	/* Store the updated SSRC back to the packet */
+	if (state->patch_ssrc)
+		rtp_hdr->ssrc = htonl(ssrc);
+
+	/* Apply the offset and store it back to the packet.
+	 * This won't change anything if the offset is 0, so the conditional is
+	 * omitted. */
+	seq += state->seq_offset;
+	rtp_hdr->sequence = htons(seq);
+	timestamp += state->timestamp_offset;
+	rtp_hdr->timestamp = htonl(timestamp);
+
+	/* Check again, whether the timestamps are still valid */
+	if (state->out_stream.ssrc == ssrc)
+		check_rtp_timestamp(endp, state, &state->out_stream, rtp_end,
+				    addr, seq, timestamp, "output",
+				    &state->out_stream.last_tsdelta);
+
+	/* Save output values */
+	state->out_stream.last_seq = seq;
+	state->out_stream.last_timestamp = timestamp;
+	state->out_stream.ssrc = ssrc;
+
+	if (payload < 0)
+		return;
+
+#if 0
+	DEBUGP(DLMGCP, "Payload hdr payload %u -> endp payload %u\n",
+	       rtp_hdr->payload_type, payload);
+	rtp_hdr->payload_type = payload;
+#endif
+}
+
+/*
+ * The below code is for dispatching. We have a dedicated port for
+ * the data coming from the net and one to discover the BTS.
+ */
+static int forward_data(int fd, struct mgcp_rtp_tap *tap, const char *buf, int len)
+{
+	if (!tap->enabled)
+		return 0;
+
+	return sendto(fd, buf, len, 0,
+		      (struct sockaddr *)&tap->forward, sizeof(tap->forward));
+}
+
+static int mgcp_send_transcoder(struct mgcp_rtp_end *end,
+				struct mgcp_config *cfg, int is_rtp,
+				const char *buf, int len)
+{
+	int rc;
+	int port;
+	struct sockaddr_in addr;
+
+	port = is_rtp ? end->rtp_port : end->rtcp_port;
+
+	addr.sin_family = AF_INET;
+	addr.sin_addr = cfg->transcoder_in;
+	addr.sin_port = port;
+
+	rc = sendto(is_rtp ?
+		end->rtp.fd :
+		end->rtcp.fd, buf, len, 0,
+		(struct sockaddr *) &addr, sizeof(addr));
+
+	if (rc != len)
+		LOGP(DLMGCP, LOGL_ERROR,
+			"Failed to send data to the transcoder: %s\n",
+			strerror(errno));
+
+	return rc;
+}
+
+int mgcp_send(struct mgcp_endpoint *endp, int dest, int is_rtp,
+	      struct sockaddr_in *addr, char *buf, int rc)
+{
+	struct mgcp_trunk_config *tcfg = endp->tcfg;
+	struct mgcp_rtp_end *rtp_end;
+	struct mgcp_rtp_state *rtp_state;
+	int tap_idx;
+
+	LOGP(DLMGCP, LOGL_DEBUG,
+	     "endpoint %x dest %s tcfg->audio_loop %d endp->conn_mode %d (== loopback: %d)\n",
+	     ENDPOINT_NUMBER(endp),
+	     dest == MGCP_DEST_NET? "net" : "bts",
+	     tcfg->audio_loop,
+	     endp->conn_mode,
+	     endp->conn_mode == MGCP_CONN_LOOPBACK);
+
+	/* For loop toggle the destination and then dispatch. */
+	if (tcfg->audio_loop)
+		dest = !dest;
+
+	/* Loop based on the conn_mode, maybe undoing the above */
+	if (endp->conn_mode == MGCP_CONN_LOOPBACK)
+		dest = !dest;
+
+	if (dest == MGCP_DEST_NET) {
+		rtp_end = &endp->net_end;
+		rtp_state = &endp->bts_state;
+		tap_idx = MGCP_TAP_NET_OUT;
+	} else {
+		rtp_end = &endp->bts_end;
+		rtp_state = &endp->net_state;
+		tap_idx = MGCP_TAP_BTS_OUT;
+	}
+	LOGP(DLMGCP, LOGL_DEBUG,
+	     "endpoint %x dest %s net_end %s %d %d bts_end %s %d %d rtp_end %s %d %d\n",
+	     ENDPOINT_NUMBER(endp),
+	     dest == MGCP_DEST_NET? "net" : "bts",
+
+	     inet_ntoa(endp->net_end.addr),
+	     ntohs(endp->net_end.rtp_port),
+	     ntohs(endp->net_end.rtcp_port),
+
+	     inet_ntoa(endp->bts_end.addr),
+	     ntohs(endp->bts_end.rtp_port),
+	     ntohs(endp->bts_end.rtcp_port),
+
+	     inet_ntoa(rtp_end->addr),
+	     ntohs(rtp_end->rtp_port),
+	     ntohs(rtp_end->rtcp_port)
+	    );
+
+	if (!rtp_end->output_enabled) {
+		rtp_end->dropped_packets += 1;
+		LOGP(DLMGCP, LOGL_DEBUG,
+		     "endpoint %x output disabled, drop to %s %s %d %d\n",
+		     ENDPOINT_NUMBER(endp),
+		     dest == MGCP_DEST_NET? "net" : "bts",
+		     inet_ntoa(rtp_end->addr),
+		     ntohs(rtp_end->rtp_port),
+		     ntohs(rtp_end->rtcp_port)
+		    );
+	} else if (is_rtp) {
+		int cont;
+		int nbytes = 0;
+		int len = rc;
+		do {
+			cont = endp->cfg->rtp_processing_cb(endp, rtp_end,
+							buf, &len, RTP_BUF_SIZE);
+			if (cont < 0)
+				break;
+
+			mgcp_patch_and_count(endp, rtp_state, rtp_end, addr, buf, len);
+		LOGP(DLMGCP, LOGL_DEBUG,
+		     "endpoint %x process/send to %s %s %d %d\n",
+		     ENDPOINT_NUMBER(endp),
+		     (dest == MGCP_DEST_NET)? "net" : "bts",
+		     inet_ntoa(rtp_end->addr),
+		     ntohs(rtp_end->rtp_port),
+		     ntohs(rtp_end->rtcp_port)
+		    );
+			forward_data(rtp_end->rtp.fd, &endp->taps[tap_idx],
+				     buf, len);
+
+			/* FIXME: HACK HACK HACK. See OS#2459.
+			 * The ip.access nano3G needs the first RTP payload's first two bytes to read hex
+			 * 'e400', or it will reject the RAB assignment. It seems to not harm other femto
+			 * cells (as long as we patch only the first RTP payload in each stream).
+			 */
+			if (tap_idx == MGCP_TAP_BTS_OUT
+			    && !rtp_state->patched_first_rtp_payload) {
+				uint8_t *data = (uint8_t*)&buf[12];
+				data[0] = 0xe4;
+				data[1] = 0x00;
+				rtp_state->patched_first_rtp_payload = true;
+			}
+
+			rc = mgcp_udp_send(rtp_end->rtp.fd,
+					   &rtp_end->addr,
+					   rtp_end->rtp_port, buf, len);
+
+			if (rc <= 0)
+				return rc;
+			nbytes += rc;
+			len = cont;
+		} while (len > 0);
+		return nbytes;
+	} else if (!tcfg->omit_rtcp) {
+		LOGP(DLMGCP, LOGL_DEBUG,
+		     "endpoint %x send to %s %s %d %d\n",
+		     ENDPOINT_NUMBER(endp),
+		     dest == MGCP_DEST_NET? "net" : "bts",
+		     inet_ntoa(rtp_end->addr),
+		     ntohs(rtp_end->rtp_port),
+		     ntohs(rtp_end->rtcp_port)
+		    );
+
+		return mgcp_udp_send(rtp_end->rtcp.fd,
+				     &rtp_end->addr,
+				     rtp_end->rtcp_port, buf, rc);
+	}
+
+	return 0;
+}
+
+static int receive_from(struct mgcp_endpoint *endp, int fd, struct sockaddr_in *addr,
+			char *buf, int bufsize)
+{
+	int rc;
+	socklen_t slen = sizeof(*addr);
+
+	rc = recvfrom(fd, buf, bufsize, 0,
+			    (struct sockaddr *) addr, &slen);
+	if (rc < 0) {
+		LOGP(DLMGCP, LOGL_ERROR, "Failed to receive message on: 0x%x errno: %d/%s\n",
+			ENDPOINT_NUMBER(endp), errno, strerror(errno));
+		return -1;
+	}
+
+	/* do not forward aynthing... maybe there is a packet from the bts */
+	if (!endp->allocated)
+		return -1;
+
+	#warning "Slight spec violation. With connection mode recvonly we should attempt to forward."
+
+	return rc;
+}
+
+static int rtp_data_net(struct osmo_fd *fd, unsigned int what)
+{
+	char buf[RTP_BUF_SIZE];
+	struct sockaddr_in addr;
+	struct mgcp_endpoint *endp;
+	int rc, proto;
+
+	endp = (struct mgcp_endpoint *) fd->data;
+
+	rc = receive_from(endp, fd->fd, &addr, buf, sizeof(buf));
+	if (rc <= 0)
+		return -1;
+
+	LOGP(DLMGCP, LOGL_DEBUG,
+	     "endpoint %x",
+	     ENDPOINT_NUMBER(endp));
+	LOGPC(DLMGCP, LOGL_DEBUG,
+	      " from net %s %d",
+	      inet_ntoa(addr.sin_addr),
+	      ntohs(addr.sin_port));
+	LOGPC(DLMGCP, LOGL_DEBUG,
+	      " net_end %s %d %d",
+	      inet_ntoa(endp->net_end.addr),
+	      ntohs(endp->net_end.rtp_port),
+	      ntohs(endp->net_end.rtcp_port));
+	LOGPC(DLMGCP, LOGL_DEBUG,
+	      " bts_end %s %d %d\n",
+	      inet_ntoa(endp->bts_end.addr),
+	      ntohs(endp->bts_end.rtp_port),
+	      ntohs(endp->bts_end.rtcp_port)
+	     );
+
+	if (memcmp(&addr.sin_addr, &endp->net_end.addr, sizeof(addr.sin_addr)) != 0) {
+		LOGP(DLMGCP, LOGL_ERROR,
+			"rtp_data_net: Endpoint 0x%x data from wrong address %s vs. ",
+			ENDPOINT_NUMBER(endp), inet_ntoa(addr.sin_addr));
+		LOGPC(DLMGCP, LOGL_ERROR,
+			"%s\n", inet_ntoa(endp->net_end.addr));
+		return -1;
+	}
+
+	switch(endp->type) {
+	case MGCP_RTP_DEFAULT:
+	case MGCP_RTP_TRANSCODED:
+		if (endp->net_end.rtp_port != addr.sin_port &&
+		    endp->net_end.rtcp_port != addr.sin_port) {
+			LOGP(DLMGCP, LOGL_ERROR,
+				"rtp_data_net: Data from wrong source port %d on 0x%x\n",
+				ntohs(addr.sin_port), ENDPOINT_NUMBER(endp));
+			return -1;
+		}
+		break;
+	case MGCP_OSMUX_BSC:
+	case MGCP_OSMUX_BSC_NAT:
+		break;
+	}
+
+	LOGP(DLMGCP, LOGL_DEBUG,
+	     "rtp_data_net: Endpoint %x data from %s %d\n",
+	     ENDPOINT_NUMBER(endp),
+	     inet_ntoa(addr.sin_addr),
+	     ntohs(addr.sin_port));
+
+	/* throw away the dummy message */
+	if (rc == 1 && buf[0] == MGCP_DUMMY_LOAD) {
+		LOGP(DLMGCP, LOGL_NOTICE, "Filtered dummy from network on 0x%x\n",
+			ENDPOINT_NUMBER(endp));
+		return 0;
+	}
+
+	proto = fd == &endp->net_end.rtp ? MGCP_PROTO_RTP : MGCP_PROTO_RTCP;
+	endp->net_end.packets += 1;
+	endp->net_end.octets += rc;
+
+	forward_data(fd->fd, &endp->taps[MGCP_TAP_NET_IN], buf, rc);
+
+	switch (endp->type) {
+	case MGCP_RTP_DEFAULT:
+		return mgcp_send(endp, MGCP_DEST_BTS, proto == MGCP_PROTO_RTP,
+				 &addr, buf, rc);
+	case MGCP_RTP_TRANSCODED:
+		return mgcp_send_transcoder(&endp->trans_net, endp->cfg,
+					    proto == MGCP_PROTO_RTP, buf, rc);
+	case MGCP_OSMUX_BSC_NAT:
+		return osmux_xfrm_to_osmux(MGCP_DEST_BTS, buf, rc, endp);
+	case MGCP_OSMUX_BSC:	/* Should not happen */
+		break;
+	}
+
+	LOGP(DLMGCP, LOGL_ERROR, "Bad MGCP type %u on endpoint %u\n",
+	     endp->type, ENDPOINT_NUMBER(endp));
+	return 0;
+}
+
+static void discover_bts(struct mgcp_endpoint *endp, int proto, struct sockaddr_in *addr)
+{
+	struct mgcp_config *cfg = endp->cfg;
+
+	if (proto == MGCP_PROTO_RTP && endp->bts_end.rtp_port == 0) {
+		if (!cfg->bts_ip ||
+		    memcmp(&addr->sin_addr,
+			   &cfg->bts_in, sizeof(cfg->bts_in)) == 0 ||
+		    memcmp(&addr->sin_addr,
+			   &endp->bts_end.addr, sizeof(endp->bts_end.addr)) == 0) {
+
+			endp->bts_end.rtp_port = addr->sin_port;
+			endp->bts_end.addr = addr->sin_addr;
+
+			LOGP(DLMGCP, LOGL_NOTICE,
+				"Found BTS for endpoint: 0x%x on port: %d/%d of %s\n",
+				ENDPOINT_NUMBER(endp), ntohs(endp->bts_end.rtp_port),
+				ntohs(endp->bts_end.rtcp_port), inet_ntoa(addr->sin_addr));
+		}
+	} else if (proto == MGCP_PROTO_RTCP && endp->bts_end.rtcp_port == 0) {
+		if (memcmp(&endp->bts_end.addr, &addr->sin_addr,
+				sizeof(endp->bts_end.addr)) == 0) {
+			endp->bts_end.rtcp_port = addr->sin_port;
+		}
+	}
+}
+
+static int rtp_data_bts(struct osmo_fd *fd, unsigned int what)
+{
+	char buf[RTP_BUF_SIZE];
+	struct sockaddr_in addr;
+	struct mgcp_endpoint *endp;
+	int rc, proto;
+
+	endp = (struct mgcp_endpoint *) fd->data;
+
+	rc = receive_from(endp, fd->fd, &addr, buf, sizeof(buf));
+	if (rc <= 0)
+		return -1;
+
+	proto = fd == &endp->bts_end.rtp ? MGCP_PROTO_RTP : MGCP_PROTO_RTCP;
+
+	/* We have no idea who called us, maybe it is the BTS. */
+	/* it was the BTS... */
+	discover_bts(endp, proto, &addr);
+
+	if (memcmp(&endp->bts_end.addr, &addr.sin_addr, sizeof(addr.sin_addr)) != 0) {
+		LOGP(DLMGCP, LOGL_ERROR,
+			"rtp_data_bts: Data from wrong bts %s on 0x%x\n",
+			inet_ntoa(addr.sin_addr), ENDPOINT_NUMBER(endp));
+		return -1;
+	}
+
+	if (endp->bts_end.rtp_port != addr.sin_port &&
+	    endp->bts_end.rtcp_port != addr.sin_port) {
+		LOGP(DLMGCP, LOGL_ERROR,
+			"rtp_data_bts: ata from wrong bts source port %d on 0x%x\n",
+			ntohs(addr.sin_port), ENDPOINT_NUMBER(endp));
+		return -1;
+	}
+
+	LOGP(DLMGCP, LOGL_DEBUG,
+	     "rtp_data_bts: Endpoint %x data from %s %d\n",
+	     ENDPOINT_NUMBER(endp),
+	     inet_ntoa(addr.sin_addr),
+	     ntohs(addr.sin_port));
+
+	/* throw away the dummy message */
+	if (rc == 1 && buf[0] == MGCP_DUMMY_LOAD) {
+		LOGP(DLMGCP, LOGL_NOTICE, "Filtered dummy from bts on 0x%x\n",
+			ENDPOINT_NUMBER(endp));
+		return 0;
+	}
+
+	/* do this before the loop handling */
+	endp->bts_end.packets += 1;
+	endp->bts_end.octets += rc;
+
+	forward_data(fd->fd, &endp->taps[MGCP_TAP_BTS_IN], buf, rc);
+
+	switch (endp->type) {
+	case MGCP_RTP_DEFAULT:
+		LOGP(DLMGCP, LOGL_DEBUG,
+		     "rtp_data_bts: Endpoint %x MGCP_RTP_DEFAULT\n",
+		     ENDPOINT_NUMBER(endp));
+		return mgcp_send(endp, MGCP_DEST_NET, proto == MGCP_PROTO_RTP,
+				 &addr, buf, rc);
+	case MGCP_RTP_TRANSCODED:
+		return mgcp_send_transcoder(&endp->trans_bts, endp->cfg,
+					    proto == MGCP_PROTO_RTP, buf, rc);
+	case MGCP_OSMUX_BSC:
+		/* OSMUX translation: BTS -> BSC */
+		return osmux_xfrm_to_osmux(MGCP_DEST_NET, buf, rc, endp);
+	case MGCP_OSMUX_BSC_NAT:
+		break;	/* Should not happen */
+	}
+
+	LOGP(DLMGCP, LOGL_ERROR, "Bad MGCP type %u on endpoint %u\n",
+	     endp->type, ENDPOINT_NUMBER(endp));
+	return 0;
+}
+
+static int rtp_data_transcoder(struct mgcp_rtp_end *end, struct mgcp_endpoint *_endp,
+			      int dest, struct osmo_fd *fd)
+{
+	char buf[RTP_BUF_SIZE];
+	struct sockaddr_in addr;
+	struct mgcp_config *cfg;
+	int rc, proto;
+
+	cfg = _endp->cfg;
+	rc = receive_from(_endp, fd->fd, &addr, buf, sizeof(buf));
+	if (rc <= 0)
+		return -1;
+
+	proto = fd == &end->rtp ? MGCP_PROTO_RTP : MGCP_PROTO_RTCP;
+
+	if (memcmp(&addr.sin_addr, &cfg->transcoder_in, sizeof(addr.sin_addr)) != 0) {
+		LOGP(DLMGCP, LOGL_ERROR,
+			"Data not coming from transcoder dest: %d %s on 0x%x\n",
+			dest, inet_ntoa(addr.sin_addr), ENDPOINT_NUMBER(_endp));
+		return -1;
+	}
+
+	if (end->rtp_port != addr.sin_port &&
+	    end->rtcp_port != addr.sin_port) {
+		LOGP(DLMGCP, LOGL_ERROR,
+			"Data from wrong transcoder dest %d source port %d on 0x%x\n",
+			dest, ntohs(addr.sin_port), ENDPOINT_NUMBER(_endp));
+		return -1;
+	}
+
+	/* throw away the dummy message */
+	if (rc == 1 && buf[0] == MGCP_DUMMY_LOAD) {
+		LOGP(DLMGCP, LOGL_NOTICE, "Filtered dummy from transcoder dest %d on 0x%x\n",
+			dest, ENDPOINT_NUMBER(_endp));
+		return 0;
+	}
+
+	end->packets += 1;
+	return mgcp_send(_endp, dest, proto == MGCP_PROTO_RTP, &addr, buf, rc);
+}
+
+static int rtp_data_trans_net(struct osmo_fd *fd, unsigned int what)
+{
+	struct mgcp_endpoint *endp;
+	endp = (struct mgcp_endpoint *) fd->data;
+
+	return rtp_data_transcoder(&endp->trans_net, endp, MGCP_DEST_NET, fd);
+}
+
+static int rtp_data_trans_bts(struct osmo_fd *fd, unsigned int what)
+{
+	struct mgcp_endpoint *endp;
+	endp = (struct mgcp_endpoint *) fd->data;
+
+	return rtp_data_transcoder(&endp->trans_bts, endp, MGCP_DEST_BTS, fd);
+}
+
+int mgcp_create_bind(const char *source_addr, struct osmo_fd *fd, int port)
+{
+	struct sockaddr_in addr;
+	int on = 1;
+
+	fd->fd = socket(AF_INET, SOCK_DGRAM, 0);
+	if (fd->fd < 0) {
+		LOGP(DLMGCP, LOGL_ERROR, "Failed to create UDP port.\n");
+		return -1;
+	}
+
+	setsockopt(fd->fd, SOL_SOCKET, SO_REUSEADDR, &on, sizeof(on));
+	memset(&addr, 0, sizeof(addr));
+	addr.sin_family = AF_INET;
+	addr.sin_port = htons(port);
+	inet_aton(source_addr, &addr.sin_addr);
+
+	if (bind(fd->fd, (struct sockaddr *) &addr, sizeof(addr)) < 0) {
+		close(fd->fd);
+		fd->fd = -1;
+		return -1;
+	}
+
+	return 0;
+}
+
+int mgcp_set_ip_tos(int fd, int tos)
+{
+	int ret;
+	ret = setsockopt(fd, IPPROTO_IP, IP_TOS,
+			 &tos, sizeof(tos));
+	return ret != 0;
+}
+
+static int bind_rtp(struct mgcp_config *cfg, const char *source_addr,
+			struct mgcp_rtp_end *rtp_end, int endpno)
+{
+	if (mgcp_create_bind(source_addr, &rtp_end->rtp,
+			     rtp_end->local_port) != 0) {
+		LOGP(DLMGCP, LOGL_ERROR, "Failed to create RTP port: %s:%d on 0x%x\n",
+		       source_addr, rtp_end->local_port, endpno);
+		goto cleanup0;
+	}
+
+	if (mgcp_create_bind(source_addr, &rtp_end->rtcp,
+			     rtp_end->local_port + 1) != 0) {
+		LOGP(DLMGCP, LOGL_ERROR, "Failed to create RTCP port: %s:%d on 0x%x\n",
+		       source_addr, rtp_end->local_port + 1, endpno);
+		goto cleanup1;
+	}
+
+	mgcp_set_ip_tos(rtp_end->rtp.fd, cfg->endp_dscp);
+	mgcp_set_ip_tos(rtp_end->rtcp.fd, cfg->endp_dscp);
+
+	rtp_end->rtp.when = BSC_FD_READ;
+	if (osmo_fd_register(&rtp_end->rtp) != 0) {
+		LOGP(DLMGCP, LOGL_ERROR, "Failed to register RTP port %d on 0x%x\n",
+			rtp_end->local_port, endpno);
+		goto cleanup2;
+	}
+
+	rtp_end->rtcp.when = BSC_FD_READ;
+	if (osmo_fd_register(&rtp_end->rtcp) != 0) {
+		LOGP(DLMGCP, LOGL_ERROR, "Failed to register RTCP port %d on 0x%x\n",
+			rtp_end->local_port + 1, endpno);
+		goto cleanup3;
+	}
+
+	return 0;
+
+cleanup3:
+	osmo_fd_unregister(&rtp_end->rtp);
+cleanup2:
+	close(rtp_end->rtcp.fd);
+	rtp_end->rtcp.fd = -1;
+cleanup1:
+	close(rtp_end->rtp.fd);
+	rtp_end->rtp.fd = -1;
+cleanup0:
+	return -1;
+}
+
+static int int_bind(const char *port,
+		    struct mgcp_rtp_end *end, int (*cb)(struct osmo_fd *, unsigned),
+		    struct mgcp_endpoint *_endp,
+		    const char *source_addr, int rtp_port)
+{
+	if (end->rtp.fd != -1 || end->rtcp.fd != -1) {
+		LOGP(DLMGCP, LOGL_ERROR, "Previous %s was still bound on %d\n",
+			port, ENDPOINT_NUMBER(_endp));
+		mgcp_free_rtp_port(end);
+	}
+
+	end->local_port = rtp_port;
+	end->rtp.cb = cb;
+	end->rtp.data = _endp;
+	end->rtcp.data = _endp;
+	end->rtcp.cb = cb;
+	return bind_rtp(_endp->cfg, source_addr, end, ENDPOINT_NUMBER(_endp));
+}
+
+int mgcp_bind_bts_rtp_port(struct mgcp_endpoint *endp, int rtp_port)
+{
+	return int_bind("bts-port", &endp->bts_end,
+			rtp_data_bts, endp,
+			mgcp_bts_src_addr(endp), rtp_port);
+}
+
+int mgcp_bind_net_rtp_port(struct mgcp_endpoint *endp, int rtp_port)
+{
+	return int_bind("net-port", &endp->net_end,
+			rtp_data_net, endp,
+			mgcp_net_src_addr(endp), rtp_port);
+}
+
+int mgcp_bind_trans_net_rtp_port(struct mgcp_endpoint *endp, int rtp_port)
+{
+	return int_bind("trans-net", &endp->trans_net,
+			rtp_data_trans_net, endp,
+			endp->cfg->source_addr, rtp_port);
+}
+
+int mgcp_bind_trans_bts_rtp_port(struct mgcp_endpoint *endp, int rtp_port)
+{
+	return int_bind("trans-bts", &endp->trans_bts,
+			rtp_data_trans_bts, endp,
+			endp->cfg->source_addr, rtp_port);
+}
+
+int mgcp_free_rtp_port(struct mgcp_rtp_end *end)
+{
+	if (end->rtp.fd != -1) {
+		close(end->rtp.fd);
+		end->rtp.fd = -1;
+		osmo_fd_unregister(&end->rtp);
+	}
+
+	if (end->rtcp.fd != -1) {
+		close(end->rtcp.fd);
+		end->rtcp.fd = -1;
+		osmo_fd_unregister(&end->rtcp);
+	}
+
+	return 0;
+}
+
+
+void mgcp_state_calc_loss(struct mgcp_rtp_state *state,
+			struct mgcp_rtp_end *end, uint32_t *expected,
+			int *loss)
+{
+	*expected = state->stats_cycles + state->stats_max_seq;
+	*expected = *expected - state->stats_base_seq + 1;
+
+	if (!state->stats_initialized) {
+		*expected = 0;
+		*loss = 0;
+		return;
+	}
+
+	/*
+	 * Make sure the sign is correct and use the biggest
+	 * positive/negative number that fits.
+	 */
+	*loss = *expected - end->packets;
+	if (*expected < end->packets) {
+		if (*loss > 0)
+			*loss = INT_MIN;
+	} else {
+		if (*loss < 0)
+			*loss = INT_MAX;
+	}
+}
+
+uint32_t mgcp_state_calc_jitter(struct mgcp_rtp_state *state)
+{
+	if (!state->stats_initialized)
+		return 0;
+	return state->stats_jitter >> 4;
+}