Merge branch 'net-next' of git://git.kernel.org/pub/scm/linux/kernel/git/vxy/lksctp-dev

Documentation/feature-removal-schedule.txt

@@ -302,18 +302,6 @@ Who:	ocfs2-devel@oss.oracle.com
 
 ---------------------------
 
-What:	SCTP_GET_PEER_ADDRS_NUM_OLD, SCTP_GET_PEER_ADDRS_OLD,
-	SCTP_GET_LOCAL_ADDRS_NUM_OLD, SCTP_GET_LOCAL_ADDRS_OLD
-When: 	June 2009
-Why:    A newer version of the options have been introduced in 2005 that
-	removes the limitions of the old API.  The sctp library has been
-        converted to use these new options at the same time.  Any user
-	space app that directly uses the old options should convert to using
-	the new options.
-Who:	Vlad Yasevich <vladislav.yasevich@hp.com>
-
----------------------------
-
 What:	Ability for non root users to shm_get hugetlb pages based on mlock
 	resource limits
 When:	2.6.31

include/linux/sctp.h

@@ -242,6 +242,7 @@ enum {
 	SCTP_DATA_FIRST_FRAG	= 0x02,
 	SCTP_DATA_NOT_FRAG	= 0x03,
 	SCTP_DATA_UNORDERED	= 0x04,
+	SCTP_DATA_SACK_IMM	= 0x08,
 };
 enum { SCTP_DATA_FRAG_MASK = 0x03, };
 

include/net/sctp/constants.h

@@ -308,6 +308,10 @@ enum { SCTP_MAX_GABS = 16 };
 
 #define SCTP_DEFAULT_MINWINDOW	1500	/* default minimum rwnd size */
 #define SCTP_DEFAULT_MAXWINDOW	65535	/* default rwnd size */
+#define SCTP_DEFAULT_RWND_SHIFT  4	/* by default, update on 1/16 of
+					 * rcvbuf, which is 1/8 of initial
+					 * window
+					 */
 #define SCTP_DEFAULT_MAXSEGMENT 1500	/* MTU size, this is the limit
                                          * to which we will raise the P-MTU.
 					 */

include/net/sctp/sm.h

@@ -243,7 +243,8 @@ struct sctp_chunk *sctp_make_op_error(const struct sctp_association *,
 				 const struct sctp_chunk *chunk,
 				 __be16 cause_code,
 				 const void *payload,
-				 size_t paylen);
+				 size_t paylen,
+				 size_t reserve_tail);
 
 struct sctp_chunk *sctp_make_asconf_update_ip(struct sctp_association *,
 					      union sctp_addr *,

include/net/sctp/structs.h

@@ -231,6 +231,11 @@ extern struct sctp_globals {
 	/* Flag to indicate whether computing and verifying checksum
 	 * is disabled. */
         int checksum_disable;
+
+	/* Threshold for rwnd update SACKS.  Receive buffer shifted this many
+	 * bits is an indicator of when to send and window update SACK.
+	 */
+	int rwnd_update_shift;
 } sctp_globals;
 
 #define sctp_rto_initial		(sctp_globals.rto_initial)
@@ -267,6 +272,7 @@ extern struct sctp_globals {
 #define sctp_prsctp_enable		(sctp_globals.prsctp_enable)
 #define sctp_auth_enable		(sctp_globals.auth_enable)
 #define sctp_checksum_disable		(sctp_globals.checksum_disable)
+#define sctp_rwnd_upd_shift		(sctp_globals.rwnd_update_shift)
 
 /* SCTP Socket type: UDP or TCP style. */
 typedef enum {
@@ -936,6 +942,8 @@ struct sctp_transport {
 	/* Data that has been sent, but not acknowledged. */
 	__u32 flight_size;
 
+	__u32 burst_limited;	/* Holds old cwnd when max.burst is applied */
+
 	/* TSN marking the fast recovery exit point */
 	__u32 fast_recovery_exit;
 
@@ -944,12 +952,6 @@ struct sctp_transport {
 	/* Source address. */
 	union sctp_addr saddr;
 
-	/* When was the last time(in jiffies) that a data packet was sent on
-	 * this transport?  This is used to adjust the cwnd when the transport
-	 * becomes inactive.
-	 */
-	unsigned long last_time_used;
-
 	/* Heartbeat interval: The endpoint sends out a Heartbeat chunk to
 	 * the destination address every heartbeat interval.
 	 */
@@ -1070,6 +1072,8 @@ void sctp_transport_put(struct sctp_transport *);
 void sctp_transport_update_rto(struct sctp_transport *, __u32);
 void sctp_transport_raise_cwnd(struct sctp_transport *, __u32, __u32);
 void sctp_transport_lower_cwnd(struct sctp_transport *, sctp_lower_cwnd_t);
+void sctp_transport_burst_limited(struct sctp_transport *);
+void sctp_transport_burst_reset(struct sctp_transport *);
 unsigned long sctp_transport_timeout(struct sctp_transport *);
 void sctp_transport_reset(struct sctp_transport *);
 void sctp_transport_update_pmtu(struct sctp_transport *, u32);

include/net/sctp/user.h

@@ -60,96 +60,49 @@ typedef __s32 sctp_assoc_t;
 /* The following symbols come from the Sockets API Extensions for
  * SCTP <draft-ietf-tsvwg-sctpsocket-07.txt>.
  */
-enum sctp_optname {
-	SCTP_RTOINFO,
-#define SCTP_RTOINFO SCTP_RTOINFO
-	SCTP_ASSOCINFO,
-#define SCTP_ASSOCINFO SCTP_ASSOCINFO
-	SCTP_INITMSG,
-#define SCTP_INITMSG SCTP_INITMSG
-	SCTP_NODELAY, 	/* Get/set nodelay option. */
-#define SCTP_NODELAY	SCTP_NODELAY
-	SCTP_AUTOCLOSE,
-#define SCTP_AUTOCLOSE SCTP_AUTOCLOSE
-	SCTP_SET_PEER_PRIMARY_ADDR, 
-#define SCTP_SET_PEER_PRIMARY_ADDR SCTP_SET_PEER_PRIMARY_ADDR
-	SCTP_PRIMARY_ADDR,
-#define SCTP_PRIMARY_ADDR SCTP_PRIMARY_ADDR
-	SCTP_ADAPTATION_LAYER,
-#define SCTP_ADAPTATION_LAYER SCTP_ADAPTATION_LAYER
-	SCTP_DISABLE_FRAGMENTS,
-#define SCTP_DISABLE_FRAGMENTS SCTP_DISABLE_FRAGMENTS
-	SCTP_PEER_ADDR_PARAMS,
-#define SCTP_PEER_ADDR_PARAMS SCTP_PEER_ADDR_PARAMS
-	SCTP_DEFAULT_SEND_PARAM,
-#define SCTP_DEFAULT_SEND_PARAM SCTP_DEFAULT_SEND_PARAM
-	SCTP_EVENTS,
-#define SCTP_EVENTS SCTP_EVENTS
-	SCTP_I_WANT_MAPPED_V4_ADDR,  /* Turn on/off mapped v4 addresses  */
-#define SCTP_I_WANT_MAPPED_V4_ADDR SCTP_I_WANT_MAPPED_V4_ADDR
-	SCTP_MAXSEG, 	/* Get/set maximum fragment. */
-#define SCTP_MAXSEG 	SCTP_MAXSEG
-	SCTP_STATUS,
-#define SCTP_STATUS SCTP_STATUS
-	SCTP_GET_PEER_ADDR_INFO,
-#define SCTP_GET_PEER_ADDR_INFO SCTP_GET_PEER_ADDR_INFO
-	SCTP_DELAYED_ACK,
-#define SCTP_DELAYED_ACK_TIME SCTP_DELAYED_ACK
-#define SCTP_DELAYED_ACK SCTP_DELAYED_ACK
-	SCTP_CONTEXT,	/* Receive Context */
-#define SCTP_CONTEXT SCTP_CONTEXT
-	SCTP_FRAGMENT_INTERLEAVE,
-#define SCTP_FRAGMENT_INTERLEAVE SCTP_FRAGMENT_INTERLEAVE
-	SCTP_PARTIAL_DELIVERY_POINT,	/* Set/Get partial delivery point */
-#define SCTP_PARTIAL_DELIVERY_POINT SCTP_PARTIAL_DELIVERY_POINT
-	SCTP_MAX_BURST,		/* Set/Get max burst */
-#define SCTP_MAX_BURST SCTP_MAX_BURST
-	SCTP_AUTH_CHUNK,	/* Set only: add a chunk type to authenticat */
-#define SCTP_AUTH_CHUNK SCTP_AUTH_CHUNK
-	SCTP_HMAC_IDENT,
-#define SCTP_HMAC_IDENT SCTP_HMAC_IDENT
-	SCTP_AUTH_KEY,
-#define SCTP_AUTH_KEY SCTP_AUTH_KEY
-	SCTP_AUTH_ACTIVE_KEY,
-#define SCTP_AUTH_ACTIVE_KEY SCTP_AUTH_ACTIVE_KEY
-	SCTP_AUTH_DELETE_KEY,
-#define SCTP_AUTH_DELETE_KEY SCTP_AUTH_DELETE_KEY
-	SCTP_PEER_AUTH_CHUNKS,		/* Read only */
-#define SCTP_PEER_AUTH_CHUNKS SCTP_PEER_AUTH_CHUNKS
-	SCTP_LOCAL_AUTH_CHUNKS,		/* Read only */
-#define SCTP_LOCAL_AUTH_CHUNKS SCTP_LOCAL_AUTH_CHUNKS
-	SCTP_GET_ASSOC_NUMBER,		/* Read only */
-#define SCTP_GET_ASSOC_NUMBER SCTP_GET_ASSOC_NUMBER
-
-
-	/* Internal Socket Options. Some of the sctp library functions are 
-	 * implemented using these socket options.
-	 */
-	SCTP_SOCKOPT_BINDX_ADD = 100,/* BINDX requests for adding addresses. */
-#define SCTP_SOCKOPT_BINDX_ADD	SCTP_SOCKOPT_BINDX_ADD
-	SCTP_SOCKOPT_BINDX_REM, /* BINDX requests for removing addresses. */
-#define SCTP_SOCKOPT_BINDX_REM	SCTP_SOCKOPT_BINDX_REM
-	SCTP_SOCKOPT_PEELOFF, 	/* peel off association. */
-#define SCTP_SOCKOPT_PEELOFF	SCTP_SOCKOPT_PEELOFF
-	SCTP_GET_PEER_ADDRS_NUM_OLD, 	/* Get number of peer addresss. */
-#define SCTP_GET_PEER_ADDRS_NUM_OLD	SCTP_GET_PEER_ADDRS_NUM_OLD
-	SCTP_GET_PEER_ADDRS_OLD, 	/* Get all peer addresss. */
-#define SCTP_GET_PEER_ADDRS_OLD	SCTP_GET_PEER_ADDRS_OLD
-	SCTP_GET_LOCAL_ADDRS_NUM_OLD, 	/* Get number of local addresss. */
-#define SCTP_GET_LOCAL_ADDRS_NUM_OLD	SCTP_GET_LOCAL_ADDRS_NUM_OLD
-	SCTP_GET_LOCAL_ADDRS_OLD, 	/* Get all local addresss. */
-#define SCTP_GET_LOCAL_ADDRS_OLD	SCTP_GET_LOCAL_ADDRS_OLD
-	SCTP_SOCKOPT_CONNECTX_OLD, /* CONNECTX old requests. */
-#define SCTP_SOCKOPT_CONNECTX_OLD	SCTP_SOCKOPT_CONNECTX_OLD
-	SCTP_GET_PEER_ADDRS, 	/* Get all peer addresss. */
-#define SCTP_GET_PEER_ADDRS	SCTP_GET_PEER_ADDRS
-	SCTP_GET_LOCAL_ADDRS, 	/* Get all local addresss. */
-#define SCTP_GET_LOCAL_ADDRS	SCTP_GET_LOCAL_ADDRS
-	SCTP_SOCKOPT_CONNECTX, /* CONNECTX requests. */
-#define SCTP_SOCKOPT_CONNECTX	SCTP_SOCKOPT_CONNECTX
-	SCTP_SOCKOPT_CONNECTX3, /* CONNECTX requests. (new implementation) */
-#define SCTP_SOCKOPT_CONNECTX3	SCTP_SOCKOPT_CONNECTX3
-};
+#define SCTP_RTOINFO	0
+#define SCTP_ASSOCINFO  1
+#define SCTP_INITMSG	2
+#define SCTP_NODELAY	3		/* Get/set nodelay option. */
+#define SCTP_AUTOCLOSE	4
+#define SCTP_SET_PEER_PRIMARY_ADDR 5
+#define SCTP_PRIMARY_ADDR	6
+#define SCTP_ADAPTATION_LAYER	7
+#define SCTP_DISABLE_FRAGMENTS	8
+#define SCTP_PEER_ADDR_PARAMS	9
+#define SCTP_DEFAULT_SEND_PARAM	10
+#define SCTP_EVENTS	11
+#define SCTP_I_WANT_MAPPED_V4_ADDR 12	/* Turn on/off mapped v4 addresses  */
+#define SCTP_MAXSEG	13		/* Get/set maximum fragment. */
+#define SCTP_STATUS	14
+#define SCTP_GET_PEER_ADDR_INFO	15
+#define SCTP_DELAYED_ACK_TIME	16
+#define SCTP_DELAYED_ACK SCTP_DELAYED_ACK_TIME
+#define SCTP_CONTEXT	17
+#define SCTP_FRAGMENT_INTERLEAVE	18
+#define SCTP_PARTIAL_DELIVERY_POINT	19 /* Set/Get partial delivery point */
+#define SCTP_MAX_BURST	20		/* Set/Get max burst */
+#define SCTP_AUTH_CHUNK	21	/* Set only: add a chunk type to authenticate */
+#define SCTP_HMAC_IDENT	22
+#define SCTP_AUTH_KEY	23
+#define SCTP_AUTH_ACTIVE_KEY	24
+#define SCTP_AUTH_DELETE_KEY	25
+#define SCTP_PEER_AUTH_CHUNKS	26	/* Read only */
+#define SCTP_LOCAL_AUTH_CHUNKS	27	/* Read only */
+#define SCTP_GET_ASSOC_NUMBER	28	/* Read only */
+
+/* Internal Socket Options. Some of the sctp library functions are
+ * implemented using these socket options.
+ */
+#define SCTP_SOCKOPT_BINDX_ADD	100	/* BINDX requests for adding addrs */
+#define SCTP_SOCKOPT_BINDX_REM	101	/* BINDX requests for removing addrs. */
+#define SCTP_SOCKOPT_PEELOFF	102	/* peel off association. */
+/* Options 104-106 are deprecated and removed. Do not use this space */
+#define SCTP_SOCKOPT_CONNECTX_OLD	107	/* CONNECTX old requests. */
+#define SCTP_GET_PEER_ADDRS	108		/* Get all peer addresss. */
+#define SCTP_GET_LOCAL_ADDRS	109		/* Get all local addresss. */
+#define SCTP_SOCKOPT_CONNECTX	110		/* CONNECTX requests. */
+#define SCTP_SOCKOPT_CONNECTX3	111	/* CONNECTX requests (updated) */
 
 /*
  * 5.2.1 SCTP Initiation Structure (SCTP_INIT)
@@ -206,6 +159,7 @@ enum sctp_sinfo_flags {
 	SCTP_UNORDERED = 1,  /* Send/receive message unordered. */
 	SCTP_ADDR_OVER = 2,  /* Override the primary destination. */
 	SCTP_ABORT=4,        /* Send an ABORT message to the peer. */
+	SCTP_SACK_IMMEDIATELY = 8,	/* SACK should be sent without delay */
 	SCTP_EOF=MSG_FIN,    /* Initiate graceful shutdown process. */	
 };
 

net/sctp/associola.c

@@ -63,6 +63,12 @@
 static void sctp_assoc_bh_rcv(struct work_struct *work);
 static void sctp_assoc_free_asconf_acks(struct sctp_association *asoc);
 
+/* Keep track of the new idr low so that we don't re-use association id
+ * numbers too fast.  It is protected by they idr spin lock is in the
+ * range of 1 - INT_MAX.
+ */
+static u32 idr_low = 1;
+
 
 /* 1st Level Abstractions. */
 
@@ -167,7 +173,7 @@ static struct sctp_association *sctp_association_init(struct sctp_association *a
 	asoc->timeouts[SCTP_EVENT_TIMEOUT_HEARTBEAT] = 0;
 	asoc->timeouts[SCTP_EVENT_TIMEOUT_SACK] = asoc->sackdelay;
 	asoc->timeouts[SCTP_EVENT_TIMEOUT_AUTOCLOSE] =
-		sp->autoclose * HZ;
+		(unsigned long)sp->autoclose * HZ;
 
 	/* Initilizes the timers */
 	for (i = SCTP_EVENT_TIMEOUT_NONE; i < SCTP_NUM_TIMEOUT_TYPES; ++i)
@@ -512,7 +518,13 @@ void sctp_assoc_set_primary(struct sctp_association *asoc,
 	 * to this destination address earlier. The sender MUST set
 	 * CYCLING_CHANGEOVER to indicate that this switch is a
 	 * double switch to the same destination address.
+	 *
+	 * Really, only bother is we have data queued or outstanding on
+	 * the association.
 	 */
+	if (!asoc->outqueue.outstanding_bytes && !asoc->outqueue.out_qlen)
+		return;
+
 	if (transport->cacc.changeover_active)
 		transport->cacc.cycling_changeover = changeover;
 
@@ -732,6 +744,7 @@ struct sctp_transport *sctp_assoc_add_peer(struct sctp_association *asoc,
 
 	peer->partial_bytes_acked = 0;
 	peer->flight_size = 0;
+	peer->burst_limited = 0;
 
 	/* Set the transport's RTO.initial value */
 	peer->rto = asoc->rto_initial;
@@ -1377,8 +1390,9 @@ static inline int sctp_peer_needs_update(struct sctp_association *asoc)
 	case SCTP_STATE_SHUTDOWN_RECEIVED:
 	case SCTP_STATE_SHUTDOWN_SENT:
 		if ((asoc->rwnd > asoc->a_rwnd) &&
-		    ((asoc->rwnd - asoc->a_rwnd) >=
-		     min_t(__u32, (asoc->base.sk->sk_rcvbuf >> 1), asoc->pathmtu)))
+		    ((asoc->rwnd - asoc->a_rwnd) >= max_t(__u32,
+			   (asoc->base.sk->sk_rcvbuf >> sctp_rwnd_upd_shift),
+			   asoc->pathmtu)))
 			return 1;
 		break;
 	default:
@@ -1545,7 +1559,12 @@ int sctp_assoc_set_id(struct sctp_association *asoc, gfp_t gfp)
 
 	spin_lock_bh(&sctp_assocs_id_lock);
 	error = idr_get_new_above(&sctp_assocs_id, (void *)asoc,
-				    1, &assoc_id);
+				    idr_low, &assoc_id);
+	if (!error) {
+		idr_low = assoc_id + 1;
+		if (idr_low == INT_MAX)
+			idr_low = 1;
+	}
 	spin_unlock_bh(&sctp_assocs_id_lock);
 	if (error == -EAGAIN)
 		goto retry;

net/sctp/chunk.c

@@ -263,9 +263,18 @@ struct sctp_datamsg *sctp_datamsg_from_user(struct sctp_association *asoc,
 		if (0 == i)
 			frag |= SCTP_DATA_FIRST_FRAG;
 
-		if ((i == (whole - 1)) && !over)
+		if ((i == (whole - 1)) && !over) {
 			frag |= SCTP_DATA_LAST_FRAG;
 
+			/* The application requests to set the I-bit of the
+			 * last DATA chunk of a user message when providing
+			 * the user message to the SCTP implementation.
+			 */
+			if ((sinfo->sinfo_flags & SCTP_EOF) ||
+			    (sinfo->sinfo_flags & SCTP_SACK_IMMEDIATELY))
+				frag |= SCTP_DATA_SACK_IMM;
+		}
+
 		chunk = sctp_make_datafrag_empty(asoc, sinfo, len, frag, 0);
 
 		if (!chunk)
@@ -297,6 +306,10 @@ struct sctp_datamsg *sctp_datamsg_from_user(struct sctp_association *asoc,
 		else
 			frag = SCTP_DATA_LAST_FRAG;
 
+		if ((sinfo->sinfo_flags & SCTP_EOF) ||
+		    (sinfo->sinfo_flags & SCTP_SACK_IMMEDIATELY))
+			frag |= SCTP_DATA_SACK_IMM;
+
 		chunk = sctp_make_datafrag_empty(asoc, sinfo, over, frag, 0);
 
 		if (!chunk)

net/sctp/output.c

@@ -429,23 +429,22 @@ int sctp_packet_transmit(struct sctp_packet *packet)
 		list_del_init(&chunk->list);
 		if (sctp_chunk_is_data(chunk)) {
 
-			if (!chunk->has_tsn) {
-				sctp_chunk_assign_ssn(chunk);
-				sctp_chunk_assign_tsn(chunk);
-
-			/* 6.3.1 C4) When data is in flight and when allowed
-			 * by rule C5, a new RTT measurement MUST be made each
-			 * round trip.  Furthermore, new RTT measurements
-			 * SHOULD be made no more than once per round-trip
-			 * for a given destination transport address.
-			 */
+			if (!chunk->resent) {
+
+				/* 6.3.1 C4) When data is in flight and when allowed
+				 * by rule C5, a new RTT measurement MUST be made each
+				 * round trip.  Furthermore, new RTT measurements
+				 * SHOULD be made no more than once per round-trip
+				 * for a given destination transport address.
+				 */
 
 				if (!tp->rto_pending) {
 					chunk->rtt_in_progress = 1;
 					tp->rto_pending = 1;
 				}
-			} else
-				chunk->resent = 1;
+			}
+
+			chunk->resent = 1;
 
 			has_data = 1;
 		}
@@ -557,8 +556,6 @@ int sctp_packet_transmit(struct sctp_packet *packet)
 		struct timer_list *timer;
 		unsigned long timeout;
 
-		tp->last_time_used = jiffies;
-
 		/* Restart the AUTOCLOSE timer when sending data. */
 		if (sctp_state(asoc, ESTABLISHED) && asoc->autoclose) {
 			timer = &asoc->timers[SCTP_EVENT_TIMEOUT_AUTOCLOSE];
@@ -617,7 +614,6 @@ static sctp_xmit_t sctp_packet_can_append_data(struct sctp_packet *packet,
 	sctp_xmit_t retval = SCTP_XMIT_OK;
 	size_t datasize, rwnd, inflight, flight_size;
 	struct sctp_transport *transport = packet->transport;
-	__u32 max_burst_bytes;
 	struct sctp_association *asoc = transport->asoc;
 	struct sctp_outq *q = &asoc->outqueue;
 
@@ -650,28 +646,6 @@ static sctp_xmit_t sctp_packet_can_append_data(struct sctp_packet *packet,
 		}
 	}
 
-	/* sctpimpguide-05 2.14.2
-	 * D) When the time comes for the sender to
-	 * transmit new DATA chunks, the protocol parameter Max.Burst MUST
-	 * first be applied to limit how many new DATA chunks may be sent.
-	 * The limit is applied by adjusting cwnd as follows:
-	 * 	if ((flightsize + Max.Burst * MTU) < cwnd)
-	 *		cwnd = flightsize + Max.Burst * MTU
-	 */
-	max_burst_bytes = asoc->max_burst * asoc->pathmtu;
-	if ((flight_size + max_burst_bytes) < transport->cwnd) {
-		transport->cwnd = flight_size + max_burst_bytes;
-		SCTP_DEBUG_PRINTK("%s: cwnd limited by max_burst: "
-				  "transport: %p, cwnd: %d, "
-				  "ssthresh: %d, flight_size: %d, "
-				  "pba: %d\n",
-				  __func__, transport,
-				  transport->cwnd,
-				  transport->ssthresh,
-				  transport->flight_size,
-				  transport->partial_bytes_acked);
-	}
-
 	/* RFC 2960 6.1  Transmission of DATA Chunks
 	 *
 	 * B) At any given time, the sender MUST NOT transmit new data
@@ -747,6 +721,8 @@ static void sctp_packet_append_data(struct sctp_packet *packet,
 	/* Has been accepted for transmission. */
 	if (!asoc->peer.prsctp_capable)
 		chunk->msg->can_abandon = 0;
+	sctp_chunk_assign_tsn(chunk);
+	sctp_chunk_assign_ssn(chunk);
 }
 
 static sctp_xmit_t sctp_packet_will_fit(struct sctp_packet *packet,

net/sctp/outqueue.c

@@ -931,6 +931,14 @@ static int sctp_outq_flush(struct sctp_outq *q, int rtx_timeout)
 				goto sctp_flush_out;
 		}
 
+		/* Apply Max.Burst limitation to the current transport in
+		 * case it will be used for new data.  We are going to
+		 * rest it before we return, but we want to apply the limit
+		 * to the currently queued data.
+		 */
+		if (transport)
+			sctp_transport_burst_limited(transport);
+
 		/* Finally, transmit new packets.  */
 		while ((chunk = sctp_outq_dequeue_data(q)) != NULL) {
 			/* RFC 2960 6.5 Every DATA chunk MUST carry a valid
@@ -976,6 +984,10 @@ static int sctp_outq_flush(struct sctp_outq *q, int rtx_timeout)
 				packet = &transport->packet;
 				sctp_packet_config(packet, vtag,
 						   asoc->peer.ecn_capable);
+				/* We've switched transports, so apply the
+				 * Burst limit to the new transport.
+				 */
+				sctp_transport_burst_limited(transport);
 			}
 
 			SCTP_DEBUG_PRINTK("sctp_outq_flush(%p, %p[%s]), ",
@@ -1011,6 +1023,13 @@ static int sctp_outq_flush(struct sctp_outq *q, int rtx_timeout)
 				break;
 
 			case SCTP_XMIT_OK:
+				/* The sender is in the SHUTDOWN-PENDING state,
+				 * The sender MAY set the I-bit in the DATA
+				 * chunk header.
+				 */
+				if (asoc->state == SCTP_STATE_SHUTDOWN_PENDING)
+					chunk->chunk_hdr->flags |= SCTP_DATA_SACK_IMM;
+
 				break;
 
 			default:
@@ -1063,6 +1082,9 @@ static int sctp_outq_flush(struct sctp_outq *q, int rtx_timeout)
 		packet = &t->packet;
 		if (!sctp_packet_empty(packet))
 			error = sctp_packet_transmit(packet);
+
+		/* Clear the burst limited state, if any */
+		sctp_transport_burst_reset(t);
 	}
 
 	return error;

net/sctp/protocol.c

@@ -1258,6 +1258,9 @@ SCTP_STATIC __init int sctp_init(void)
 	/* Set SCOPE policy to enabled */
 	sctp_scope_policy = SCTP_SCOPE_POLICY_ENABLE;
 
+	/* Set the default rwnd update threshold */
+	sctp_rwnd_upd_shift		= SCTP_DEFAULT_RWND_SHIFT;
+
 	sctp_sysctl_register();
 
 	INIT_LIST_HEAD(&sctp_address_families);

net/sctp/sm_make_chunk.c

@@ -987,7 +987,10 @@ static void *sctp_addto_param(struct sctp_chunk *chunk, int len,
 
 	target = skb_put(chunk->skb, len);
 
-	memcpy(target, data, len);
+	if (data)
+		memcpy(target, data, len);
+	else
+		memset(target, 0, len);
 
 	/* Adjust the chunk length field.  */
 	chunk->chunk_hdr->length = htons(chunklen + len);
@@ -1129,16 +1132,18 @@ static struct sctp_chunk *sctp_make_op_error_space(
 struct sctp_chunk *sctp_make_op_error(const struct sctp_association *asoc,
 				 const struct sctp_chunk *chunk,
 				 __be16 cause_code, const void *payload,
-				 size_t paylen)
+				 size_t paylen, size_t reserve_tail)
 {
 	struct sctp_chunk *retval;
 
-	retval = sctp_make_op_error_space(asoc, chunk, paylen);
+	retval = sctp_make_op_error_space(asoc, chunk, paylen + reserve_tail);
 	if (!retval)
 		goto nodata;
 
-	sctp_init_cause(retval, cause_code, paylen);
+	sctp_init_cause(retval, cause_code, paylen + reserve_tail);
 	sctp_addto_chunk(retval, paylen, payload);
+	if (reserve_tail)
+		sctp_addto_param(retval, reserve_tail, NULL);
 
 nodata:
 	return retval;

net/sctp/sm_sideeffect.c

@@ -217,8 +217,7 @@ static int sctp_gen_sack(struct sctp_association *asoc, int force,
 		sctp_add_cmd_sf(commands, SCTP_CMD_TIMER_RESTART,
 				SCTP_TO(SCTP_EVENT_TIMEOUT_SACK));
 	} else {
-		if (asoc->a_rwnd > asoc->rwnd)
-			asoc->a_rwnd = asoc->rwnd;
+		asoc->a_rwnd = asoc->rwnd;
 		sack = sctp_make_sack(asoc);
 		if (!sack)
 			goto nomem;
@@ -1418,6 +1417,8 @@ static int sctp_cmd_interpreter(sctp_event_t event_type,
 			asoc->init_last_sent_to = t;
 			chunk->transport = t;
 			t->init_sent_count++;
+			/* Set the new transport as primary */
+			sctp_assoc_set_primary(asoc, t);
 			break;
 
 		case SCTP_CMD_INIT_RESTART:

net/sctp/sm_statefuns.c

@@ -996,14 +996,15 @@ sctp_disposition_t sctp_sf_sendbeat_8_3(const struct sctp_endpoint *ep,
 				sctp_sf_heartbeat(ep, asoc, type, arg,
 						  commands))
 			return SCTP_DISPOSITION_NOMEM;
+
 		/* Set transport error counter and association error counter
 		 * when sending heartbeat.
 		 */
-		sctp_add_cmd_sf(commands, SCTP_CMD_TRANSPORT_IDLE,
-				SCTP_TRANSPORT(transport));
 		sctp_add_cmd_sf(commands, SCTP_CMD_TRANSPORT_HB_SENT,
 				SCTP_TRANSPORT(transport));
 	}
+	sctp_add_cmd_sf(commands, SCTP_CMD_TRANSPORT_IDLE,
+			SCTP_TRANSPORT(transport));
 	sctp_add_cmd_sf(commands, SCTP_CMD_HB_TIMER_UPDATE,
 			SCTP_TRANSPORT(transport));
 
@@ -1720,7 +1721,7 @@ static sctp_disposition_t sctp_sf_do_dupcook_a(const struct sctp_endpoint *ep,
 
 		err = sctp_make_op_error(asoc, chunk,
 					 SCTP_ERROR_COOKIE_IN_SHUTDOWN,
-					 NULL, 0);
+					 NULL, 0, 0);
 		if (err)
 			sctp_add_cmd_sf(commands, SCTP_CMD_REPLY,
 					SCTP_CHUNK(err));
@@ -2868,6 +2869,7 @@ sctp_disposition_t sctp_sf_eat_data_6_2(const struct sctp_endpoint *ep,
 					sctp_cmd_seq_t *commands)
 {
 	struct sctp_chunk *chunk = arg;
+	sctp_arg_t force = SCTP_NOFORCE();
 	int error;
 
 	if (!sctp_vtag_verify(chunk, asoc)) {
@@ -2901,6 +2903,9 @@ sctp_disposition_t sctp_sf_eat_data_6_2(const struct sctp_endpoint *ep,
 		BUG();
 	}
 
+	if (chunk->chunk_hdr->flags & SCTP_DATA_SACK_IMM)
+		force = SCTP_FORCE();
+
 	if (asoc->autoclose) {
 		sctp_add_cmd_sf(commands, SCTP_CMD_TIMER_RESTART,
 				SCTP_TO(SCTP_EVENT_TIMEOUT_AUTOCLOSE));
@@ -2929,7 +2934,7 @@ sctp_disposition_t sctp_sf_eat_data_6_2(const struct sctp_endpoint *ep,
 	 * more aggressive than the following algorithms allow.
 	 */
 	if (chunk->end_of_packet)
-		sctp_add_cmd_sf(commands, SCTP_CMD_GEN_SACK, SCTP_NOFORCE());
+		sctp_add_cmd_sf(commands, SCTP_CMD_GEN_SACK, force);
 
 	return SCTP_DISPOSITION_CONSUME;
 
@@ -2954,7 +2959,7 @@ sctp_disposition_t sctp_sf_eat_data_6_2(const struct sctp_endpoint *ep,
 
 discard_noforce:
 	if (chunk->end_of_packet)
-		sctp_add_cmd_sf(commands, SCTP_CMD_GEN_SACK, SCTP_NOFORCE());
+		sctp_add_cmd_sf(commands, SCTP_CMD_GEN_SACK, force);
 
 	return SCTP_DISPOSITION_DISCARD;
 consume:
@@ -3973,7 +3978,7 @@ sctp_disposition_t sctp_sf_eat_auth(const struct sctp_endpoint *ep,
 			err_chunk = sctp_make_op_error(asoc, chunk,
 							SCTP_ERROR_UNSUP_HMAC,
 							&auth_hdr->hmac_id,
-							sizeof(__u16));
+							sizeof(__u16), 0);
 			if (err_chunk) {
 				sctp_add_cmd_sf(commands, SCTP_CMD_REPLY,
 						SCTP_CHUNK(err_chunk));
@@ -4065,7 +4070,8 @@ sctp_disposition_t sctp_sf_unk_chunk(const struct sctp_endpoint *ep,
 		hdr = unk_chunk->chunk_hdr;
 		err_chunk = sctp_make_op_error(asoc, unk_chunk,
 					       SCTP_ERROR_UNKNOWN_CHUNK, hdr,
-					       WORD_ROUND(ntohs(hdr->length)));
+					       WORD_ROUND(ntohs(hdr->length)),
+					       0);
 		if (err_chunk) {
 			sctp_add_cmd_sf(commands, SCTP_CMD_REPLY,
 					SCTP_CHUNK(err_chunk));
@@ -4084,7 +4090,8 @@ sctp_disposition_t sctp_sf_unk_chunk(const struct sctp_endpoint *ep,
 		hdr = unk_chunk->chunk_hdr;
 		err_chunk = sctp_make_op_error(asoc, unk_chunk,
 					       SCTP_ERROR_UNKNOWN_CHUNK, hdr,
-					       WORD_ROUND(ntohs(hdr->length)));
+					       WORD_ROUND(ntohs(hdr->length)),
+					       0);
 		if (err_chunk) {
 			sctp_add_cmd_sf(commands, SCTP_CMD_REPLY,
 					SCTP_CHUNK(err_chunk));
@@ -6048,7 +6055,8 @@ static int sctp_eat_data(const struct sctp_association *asoc,
 
 		err = sctp_make_op_error(asoc, chunk, SCTP_ERROR_INV_STRM,
 					 &data_hdr->stream,
-					 sizeof(data_hdr->stream));
+					 sizeof(data_hdr->stream),
+					 sizeof(u16));
 		if (err)
 			sctp_add_cmd_sf(commands, SCTP_CMD_REPLY,
 					SCTP_CHUNK(err));

net/sctp/sysctl.c

@@ -52,6 +52,7 @@ static int int_max = INT_MAX;
 static int sack_timer_min = 1;
 static int sack_timer_max = 500;
 static int addr_scope_max = 3; /* check sctp_scope_policy_t in include/net/sctp/constants.h for max entries */
+static int rwnd_scale_max = 16;
 
 extern int sysctl_sctp_mem[3];
 extern int sysctl_sctp_rmem[3];
@@ -284,6 +285,18 @@ static ctl_table sctp_table[] = {
 		.extra1		= &zero,
 		.extra2		= &addr_scope_max,
 	},
+	{
+		.ctl_name	= CTL_UNNUMBERED,
+		.procname	= "rwnd_update_shift",
+		.data		= &sctp_rwnd_upd_shift,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= &proc_dointvec_minmax,
+		.strategy	= &sysctl_intvec,
+		.extra1		= &one,
+		.extra2		= &rwnd_scale_max,
+	},
+
 	{ .ctl_name = 0 }
 };
 

net/sctp/transport.c

@@ -83,7 +83,6 @@ static struct sctp_transport *sctp_transport_init(struct sctp_transport *peer,
 	peer->fast_recovery = 0;
 
 	peer->last_time_heard = jiffies;
-	peer->last_time_used = jiffies;
 	peer->last_time_ecne_reduced = jiffies;
 
 	peer->init_sent_count = 0;
@@ -565,10 +564,8 @@ void sctp_transport_lower_cwnd(struct sctp_transport *transport,
 		 * to be done every RTO interval, we do it every hearbeat
 		 * interval.
 		 */
-		if (time_after(jiffies, transport->last_time_used +
-					transport->rto))
-			transport->cwnd = max(transport->cwnd/2,
-						 4*transport->asoc->pathmtu);
+		transport->cwnd = max(transport->cwnd/2,
+					 4*transport->asoc->pathmtu);
 		break;
 	}
 
@@ -579,6 +576,43 @@ void sctp_transport_lower_cwnd(struct sctp_transport *transport,
 			  transport->cwnd, transport->ssthresh);
 }
 
+/* Apply Max.Burst limit to the congestion window:
+ * sctpimpguide-05 2.14.2
+ * D) When the time comes for the sender to
+ * transmit new DATA chunks, the protocol parameter Max.Burst MUST
+ * first be applied to limit how many new DATA chunks may be sent.
+ * The limit is applied by adjusting cwnd as follows:
+ * 	if ((flightsize+ Max.Burst * MTU) < cwnd)
+ * 		cwnd = flightsize + Max.Burst * MTU
+ */
+
+void sctp_transport_burst_limited(struct sctp_transport *t)
+{
+	struct sctp_association *asoc = t->asoc;
+	u32 old_cwnd = t->cwnd;
+	u32 max_burst_bytes;
+
+	if (t->burst_limited)
+		return;
+
+	max_burst_bytes = t->flight_size + (asoc->max_burst * asoc->pathmtu);
+	if (max_burst_bytes < old_cwnd) {
+		t->cwnd = max_burst_bytes;
+		t->burst_limited = old_cwnd;
+	}
+}
+
+/* Restore the old cwnd congestion window, after the burst had it's
+ * desired effect.
+ */
+void sctp_transport_burst_reset(struct sctp_transport *t)
+{
+	if (t->burst_limited) {
+		t->cwnd = t->burst_limited;
+		t->burst_limited = 0;
+	}
+}
+
 /* What is the next timeout value for this transport? */
 unsigned long sctp_transport_timeout(struct sctp_transport *t)
 {
@@ -601,6 +635,7 @@ void sctp_transport_reset(struct sctp_transport *t)
 	 * (see Section 6.2.1)
 	 */
 	t->cwnd = min(4*asoc->pathmtu, max_t(__u32, 2*asoc->pathmtu, 4380));
+	t->burst_limited = 0;
 	t->ssthresh = asoc->peer.i.a_rwnd;
 	t->last_rto = t->rto = asoc->rto_initial;
 	t->rtt = 0;