Merge http://linux-lksctp.bkbits.net/lksctp-2.5.work

into touki.austin.ibm.com:/home/jgrimm/bk/lksctp-2.5.work

include/net/sctp/structs.h

@@ -450,7 +450,7 @@ static inline __u16 sctp_ssn_next(struct sctp_stream *stream, __u16 id)
 }
 
 /* Structure to track chunk fragments that have been acked, but peer
- * fragments of the same message have not.  
+ * fragments of the same message have not.
  */
 struct sctp_datamsg {
 	/* Chunks waiting to be submitted to lower layer. */
@@ -459,9 +459,13 @@ struct sctp_datamsg {
 	struct list_head track;
 	/* Reference counting. */
 	atomic_t refcnt;
+	/* When is this message no longer interesting to the peer? */
+	unsigned long expires_at; 
 	/* Did the messenge fail to send? */
 	int send_error;
 	char send_failed;
+	/* Control whether fragments from this message can expire. */
+	char can_expire;
 };
 
 struct sctp_datamsg *sctp_datamsg_from_user(struct sctp_association *,
@@ -473,6 +477,9 @@ void sctp_datamsg_hold(struct sctp_datamsg *);
 void sctp_datamsg_free(struct sctp_datamsg *);
 void sctp_datamsg_track(struct sctp_chunk *);
 void sctp_datamsg_assign(struct sctp_datamsg *, struct sctp_chunk *);
+void sctp_datamsg_fail(struct sctp_chunk *, int error);
+int sctp_datamsg_expires(struct sctp_chunk *);
+
 
 /* RFC2960 1.4 Key Terms
  *
@@ -545,18 +552,6 @@ struct sctp_chunk {
 	/* We fill this in if we are calculating RTT. */
 	unsigned long sent_at;
 
-	__u8 rtt_in_progress;  /* Is this chunk used for RTT calculation? */
-	__u8 resent;           /* Has this chunk ever been retransmitted. */
-	__u8 has_tsn;          /* Does this chunk have a TSN yet? */
-	__u8 has_ssn;          /* Does this chunk have a SSN yet? */
-	__u8 singleton;        /* Was this the only chunk in the packet? */
-	__u8 end_of_packet;    /* Was this the last chunk in the packet? */
-	__u8 ecn_ce_done;      /* Have we processed the ECN CE bit? */
-	__u8 pdiscard;	       /* Discard the whole packet now? */
-	__u8 tsn_gap_acked;	  /* Is this chunk acked by a GAP ACK? */
-	__u8 fast_retransmit;    /* Is this chunk fast retransmitted? */
-	__u8 tsn_missing_report; /* Data chunk missing counter. */
-
 	/* What is the origin IP address for this chunk?  */
 	union sctp_addr source;
 	/* Destination address for this chunk. */
@@ -570,6 +565,18 @@ struct sctp_chunk {
 	 * go.  It is NULL if we have no preference.
 	 */
 	struct sctp_transport *transport;
+
+	__u8 rtt_in_progress;	/* Is this chunk used for RTT calculation? */
+	__u8 resent;		/* Has this chunk ever been retransmitted. */
+	__u8 has_tsn;		/* Does this chunk have a TSN yet? */
+	__u8 has_ssn;		/* Does this chunk have a SSN yet? */
+	__u8 singleton;		/* Was this the only chunk in the packet? */
+	__u8 end_of_packet;	/* Was this the last chunk in the packet? */
+	__u8 ecn_ce_done;	/* Have we processed the ECN CE bit? */
+	__u8 pdiscard;		/* Discard the whole packet now? */
+	__u8 tsn_gap_acked;	/* Is this chunk acked by a GAP ACK? */
+	__u8 fast_retransmit;	 /* Is this chunk fast retransmitted? */
+	__u8 tsn_missing_report; /* Data chunk missing counter. */
 };
 
 void sctp_chunk_hold(struct sctp_chunk *);
@@ -580,10 +587,10 @@ struct sctp_chunk *sctp_make_chunk(const struct sctp_association *, __u8 type,
 				   __u8 flags, int size);
 void sctp_chunk_free(struct sctp_chunk *);
 void  *sctp_addto_chunk(struct sctp_chunk *, int len, const void *data);
-struct sctp_chunk *sctp_chunkify(struct sk_buff *, 
+struct sctp_chunk *sctp_chunkify(struct sk_buff *,
 				 const struct sctp_association *,
 				 struct sock *);
-void sctp_init_addrs(struct sctp_chunk *, union sctp_addr *, 
+void sctp_init_addrs(struct sctp_chunk *, union sctp_addr *,
 		     union sctp_addr *);
 const union sctp_addr *sctp_source(const struct sctp_chunk *chunk);
 
@@ -847,7 +854,7 @@ struct sctp_transport {
 
 		/* A flag which indicates the occurrence of a changeover */
 		char changeover_active;
-	
+
 		/* A glag which indicates whether the change of primary is
 		 * the first switch to this destination address during an
 		 * active switch.
@@ -1024,7 +1031,7 @@ struct sctp_bind_addr {
 struct sctp_bind_addr *sctp_bind_addr_new(int gfp_mask);
 void sctp_bind_addr_init(struct sctp_bind_addr *, __u16 port);
 void sctp_bind_addr_free(struct sctp_bind_addr *);
-int sctp_bind_addr_copy(struct sctp_bind_addr *dest, 
+int sctp_bind_addr_copy(struct sctp_bind_addr *dest,
 			const struct sctp_bind_addr *src,
 			sctp_scope_t scope, int gfp,int flags);
 int sctp_add_bind_addr(struct sctp_bind_addr *, union sctp_addr *,
@@ -1613,7 +1620,7 @@ struct sctp_association {
 	/* Need to send an ECNE Chunk? */
 	char need_ecne;
 
-	/* Is it a temporary association? */ 
+	/* Is it a temporary association? */
 	char temp;
 };
 

net/sctp/chunk.c

@@ -55,8 +55,8 @@ void sctp_datamsg_init(struct sctp_datamsg *msg)
 	atomic_set(&msg->refcnt, 1);
 	msg->send_failed = 0;
 	msg->send_error = 0;
+	msg->can_expire = 0;
 	INIT_LIST_HEAD(&msg->chunks);
-	INIT_LIST_HEAD(&msg->track);
 }
 
 /* Allocate and initialize datamsg. */
@@ -84,7 +84,7 @@ static void sctp_datamsg_destroy(struct sctp_datamsg *msg)
 	notify = msg->send_failed ? -1 : 0;
 
 	/* Release all references. */
-	list_for_each_safe(pos, temp, &msg->track) {
+	list_for_each_safe(pos, temp, &msg->chunks) {
 		list_del(pos);
 		chunk = list_entry(pos, struct sctp_chunk, frag_list);
 		/* Check whether we _really_ need to notify. */
@@ -94,7 +94,7 @@ static void sctp_datamsg_destroy(struct sctp_datamsg *msg)
 				error = msg->send_error;
 			else
 				error = asoc->outqueue.error;
-				
+
 			sp = sctp_sk(asoc->base.sk);
 			notify = sctp_ulpevent_type_enabled(SCTP_SEND_FAILED,
 							    &sp->subscribe);
@@ -107,8 +107,8 @@ static void sctp_datamsg_destroy(struct sctp_datamsg *msg)
 				sent = SCTP_DATA_SENT;
 			else
 				sent = SCTP_DATA_UNSENT;
-		       
-			ev = sctp_ulpevent_make_send_failed(asoc, chunk, sent, 
+
+			ev = sctp_ulpevent_make_send_failed(asoc, chunk, sent,
 							    error, GFP_ATOMIC);
 			if (ev)
 				sctp_ulpq_tail_event(&asoc->ulpq, ev);
@@ -146,7 +146,6 @@ void sctp_datamsg_free(struct sctp_datamsg *msg)
 void sctp_datamsg_track(struct sctp_chunk *chunk)
 {
 	sctp_chunk_hold(chunk);
-	list_add_tail(&chunk->frag_list, &chunk->msg->track);
 }
 
 /* Assign a chunk to this datamsg. */
@@ -179,6 +178,20 @@ struct sctp_datamsg *sctp_datamsg_from_user(struct sctp_association *asoc,
 	if (!msg)
 		return NULL;
 
+	/* Note: Calculate this outside of the loop, so that all fragments
+	 * have the same expiration.
+	 */
+	if (sinfo->sinfo_timetolive) {
+		struct timeval tv;
+		__u32 ttl = sinfo->sinfo_timetolive;
+
+		/* sinfo_timetolive is in milliseconds */
+		tv.tv_sec = ttl / 1000;
+		tv.tv_usec = ttl % 1000 * 1000;
+		msg->expires_at = jiffies + timeval_to_jiffies(&tv);
+		msg->can_expire = 1;
+	}
+
 	/* What is a reasonable fragmentation point right now? */
 	max = asoc->pmtu;
 	if (max < SCTP_MIN_PMTU)
@@ -191,7 +204,6 @@ struct sctp_datamsg *sctp_datamsg_from_user(struct sctp_association *asoc,
 
 	/* Subtract out the overhead of a data chunk header. */
 	max -= sizeof(struct sctp_data_chunk);
-
 	whole = 0;
 
 	/* If user has specified smaller fragmentation, make it so. */
@@ -289,3 +301,27 @@ struct sctp_datamsg *sctp_datamsg_from_user(struct sctp_association *asoc,
 	sctp_datamsg_free(msg);
 	return NULL;
 }
+
+/* Check whether this message has expired. */
+int sctp_datamsg_expires(struct sctp_chunk *chunk)
+{
+	struct sctp_datamsg *msg = chunk->msg;
+
+	/* FIXME: When PR-SCTP is supported we can make this
+	 * check more lenient.
+	 */
+	if (!msg->can_expire)
+		return 0;
+
+	if (time_after(jiffies, msg->expires_at))
+		return 1;
+
+	return 0;
+}
+
+/* This chunk (and consequently entire message) has failed in its sending. */
+void sctp_datamsg_fail(struct sctp_chunk *chunk, int error)
+{
+	chunk->msg->send_failed = 1;
+	chunk->msg->send_error = error;
+}

net/sctp/output.c

@@ -356,6 +356,7 @@ int sctp_packet_transmit(struct sctp_packet *packet)
 		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
@@ -627,6 +628,8 @@ static sctp_xmit_t sctp_packet_append_data(struct sctp_packet *packet,
 		rwnd = 0;
 
 	asoc->peer.rwnd = rwnd;
+	/* Has been accepted for transmission. */
+	chunk->msg->can_expire = 0;
 
 finish:
 	return retval;

net/sctp/outqueue.c

@@ -165,7 +165,7 @@ static inline int sctp_cacc_skip_3_1(struct sctp_transport *primary,
  */
 static inline int sctp_cacc_skip_3_2(struct sctp_transport *primary, __u32 tsn)
 {
-	if (primary->cacc.cycling_changeover && 
+	if (primary->cacc.cycling_changeover &&
 	    TSN_lt(tsn, primary->cacc.next_tsn_at_change))
 		return 1;
 	return 0;
@@ -174,7 +174,7 @@ static inline int sctp_cacc_skip_3_2(struct sctp_transport *primary, __u32 tsn)
 /*
  * SFR-CACC algorithm:
  * 3) If the missing report count for TSN t is to be
- * incremented according to [RFC2960] and 
+ * incremented according to [RFC2960] and
  * [SCTP_STEWART-2002], and CHANGEOVER_ACTIVE is set,
  * then the sender MUST futher execute steps 3.1 and
  * 3.2 to determine if the missing report count for
@@ -251,8 +251,7 @@ void sctp_outq_teardown(struct sctp_outq *q)
 			chunk = list_entry(lchunk, struct sctp_chunk,
 					   transmitted_list);
 			/* Mark as part of a failed message. */
-			chunk->msg->send_failed = 1;
-			/* Generate a SEND FAILED event. */
+			sctp_datamsg_fail(chunk, q->error);
 			sctp_chunk_free(chunk);
 		}
 	}
@@ -262,6 +261,7 @@ void sctp_outq_teardown(struct sctp_outq *q)
 		list_del(lchunk);
 		chunk = list_entry(lchunk, struct sctp_chunk,
 				   transmitted_list);
+		sctp_datamsg_fail(chunk, q->error);
 		sctp_chunk_free(chunk);
 	}
 
@@ -270,6 +270,7 @@ void sctp_outq_teardown(struct sctp_outq *q)
 		list_del(lchunk);
 		chunk = list_entry(lchunk, struct sctp_chunk,
 				   transmitted_list);
+		sctp_datamsg_fail(chunk, q->error);
 		sctp_chunk_free(chunk);
 	}
 
@@ -277,7 +278,7 @@ void sctp_outq_teardown(struct sctp_outq *q)
 	while ((chunk = sctp_outq_dequeue_data(q))) {
 
 		/* Mark as send failure. */
-		chunk->msg->send_failed = 1;
+		sctp_datamsg_fail(chunk, q->error);
 		sctp_chunk_free(chunk);
 	}
 
@@ -800,27 +801,25 @@ int sctp_outq_flush(struct sctp_outq *q, int rtx_timeout)
 			 */
 			if (chunk->sinfo.sinfo_stream >=
 			    asoc->c.sinit_num_ostreams) {
-				/* Mark as failed send. */
-				chunk->msg->send_failed = 1;
-				chunk->msg->send_error = SCTP_ERROR_INV_STRM;
 
-				/* Free the chunk. */
+				/* Mark as s failed send. */
+				sctp_datamsg_fail(chunk, SCTP_ERROR_INV_STRM);
 				sctp_chunk_free(chunk);
 				continue;
 			}
 
-			/* Now do delayed assignment of SSN.  This will
-			 * probably change again when we start supporting
-			 * large (> approximately 2^16) size messages.
-			 */
-			sctp_chunk_assign_ssn(chunk);
+			/* Has this chunk expired? */
+			if (sctp_datamsg_expires(chunk)) {
+				sctp_datamsg_fail(chunk, 0);
+				sctp_chunk_free(chunk);
+				continue;
+			}
 
 			/* If there is a specified transport, use it.
 			 * Otherwise, we want to use the active path.
 			 */
 			new_transport = chunk->transport;
-			if (new_transport == NULL ||
-			    !new_transport->active)
+			if (!new_transport || !new_transport->active)
 				new_transport = asoc->peer.active_path;
 
 			/* Change packets if necessary.  */
@@ -1024,10 +1023,10 @@ int sctp_outq_sack(struct sctp_outq *q, struct sctp_sackhdr *sack)
 
 	/*
 	 * SFR-CACC algorithm:
-	 * On receipt of a SACK the sender SHOULD execute the 
+	 * On receipt of a SACK the sender SHOULD execute the
 	 * following statements.
 	 *
-	 * 1) If the cumulative ack in the SACK passes next tsn_at_change 
+	 * 1) If the cumulative ack in the SACK passes next tsn_at_change
 	 * on the current primary, the CHANGEOVER_ACTIVE flag SHOULD be
 	 * cleared. The CYCLING_CHANGEOVER flag SHOULD also be cleared for
 	 * all destinations.
@@ -1041,7 +1040,7 @@ int sctp_outq_sack(struct sctp_outq *q, struct sctp_sackhdr *sack)
 		}
 	}
 
-	/* 
+	/*
 	 * SFR-CACC algorithm:
 	 * 2) If the SACK contains gap acks and the flag CHANGEOVER_ACTIVE
 	 * is set the receiver of the SACK MUST take the following actions:
@@ -1087,7 +1086,7 @@ int sctp_outq_sack(struct sctp_outq *q, struct sctp_sackhdr *sack)
 				       transport, sack, highest_new_tsn);
 		/*
 		 * SFR-CACC algorithm:
-		 * C) Let count_of_newacks be the number of 
+		 * C) Let count_of_newacks be the number of
 		 * destinations for which cacc_saw_newack is set.
 		 */
 		if (transport->cacc.cacc_saw_newack)
@@ -1097,7 +1096,7 @@ int sctp_outq_sack(struct sctp_outq *q, struct sctp_sackhdr *sack)
 	list_for_each(pos, transport_list) {
 		transport  = list_entry(pos, struct sctp_transport,
 					transports);
-		sctp_mark_missing(q, &transport->transmitted, transport, 
+		sctp_mark_missing(q, &transport->transmitted, transport,
 				  highest_new_tsn, count_of_newacks);
 	}
 
@@ -1275,8 +1274,8 @@ static void sctp_check_transmitted(struct sctp_outq *q,
 					 * the destination that the TSN was
 					 * sent to.
 					 */
-					if (transport && 
-					    sack->num_gap_ack_blocks && 
+					if (transport &&
+					    sack->num_gap_ack_blocks &&
 					    q->asoc->peer.primary_path->cacc.
 					    changeover_active)
 						transport->cacc.cacc_saw_newack
@@ -1529,12 +1528,12 @@ static void sctp_mark_missing(struct sctp_outq *q,
 		    TSN_lt(tsn, highest_new_tsn_in_sack)) {
 
 			/* SFR-CACC may require us to skip marking
-			 * this chunk as missing. 
+			 * this chunk as missing.
 			 */
 			if (!transport || !sctp_cacc_skip(primary, transport,
 					    count_of_newacks, tsn)) {
 				chunk->tsn_missing_report++;
-				
+
 				SCTP_DEBUG_PRINTK(
 					"%s: TSN 0x%x missing counter: %d\n",
 					__FUNCTION__, tsn,

net/sctp/sm_make_chunk.c

@@ -992,7 +992,7 @@ struct sctp_chunk *sctp_chunkify(struct sk_buff *skb,
 	retval->has_tsn		= 0;
 	retval->has_ssn         = 0;
 	retval->rtt_in_progress	= 0;
-	retval->sent_at	= jiffies;
+	retval->sent_at		= 0;
 	retval->singleton	= 1;
 	retval->end_of_packet	= 0;
 	retval->ecn_ce_done	= 0;

net/sctp/sm_statefuns.c

@@ -2372,7 +2372,8 @@ sctp_disposition_t sctp_sf_eat_data_6_2(const struct sctp_endpoint *ep,
 	 * PMTU.  In cases, such as loopback, this might be a rather
 	 * large spill over.
 	 */
-	if (asoc->rwnd_over || (datalen > asoc->rwnd + asoc->frag_point)) {
+	if (!asoc->rwnd || asoc->rwnd_over || 
+	    (datalen > asoc->rwnd + asoc->frag_point)) {
 
 		/* If this is the next TSN, consider reneging to make
 		 * room.   Note: Playing nice with a confused sender.  A

net/sctp/socket.c

@@ -826,7 +826,7 @@ SCTP_STATIC int sctp_sendmsg(struct kiocb *iocb, struct sock *sk,
 	long timeo;
 	__u16 sinfo_flags = 0;
 	struct sctp_datamsg *datamsg;
-	struct list_head *pos, *temp;
+	struct list_head *pos;
 	int msg_flags = msg->msg_flags;
 
 	SCTP_DEBUG_PRINTK("sctp_sendmsg(sk: %p, msg: %p, msg_len: %d)\n",
@@ -1133,9 +1133,8 @@ SCTP_STATIC int sctp_sendmsg(struct kiocb *iocb, struct sock *sk,
 	}
 
 	/* Now send the (possibly) fragmented message. */
-	list_for_each_safe(pos, temp, &datamsg->chunks) {
+	list_for_each(pos, &datamsg->chunks) {
 		chunk = list_entry(pos, struct sctp_chunk, frag_list);
-		list_del_init(pos);
 		sctp_datamsg_track(chunk);
 
 		/* Do accounting for the write space.  */
@@ -1143,7 +1142,11 @@ SCTP_STATIC int sctp_sendmsg(struct kiocb *iocb, struct sock *sk,
 
 		chunk->transport = chunk_tp;
 
-		/* Send it to the lower layers.  */
+		/* Send it to the lower layers.  Note:  all chunks
+		 * must either fail or succeed.   The lower layer
+		 * works that way today.  Keep it that way or this
+		 * breaks.
+		 */
 		err = sctp_primitive_SEND(asoc, chunk);
 		/* Did the lower layer accept the chunk? */
 		if (err)
@@ -1152,7 +1155,7 @@ SCTP_STATIC int sctp_sendmsg(struct kiocb *iocb, struct sock *sk,
 	}
 
 	sctp_datamsg_free(datamsg);
-	if (err) 
+	if (err)
 		goto out_free;
 	else
 		err = msg_len;
@@ -3815,7 +3818,7 @@ static void sctp_sock_migrate(struct sock *oldsk, struct sock *newsk,
 
 	/* If the association on the newsk is already closed before accept()
 	 * is called, set RCV_SHUTDOWN flag.
-	 */ 
+	 */
 	if (sctp_state(assoc, CLOSED) && sctp_style(newsk, TCP))
 		newsk->shutdown |= RCV_SHUTDOWN;
 

net/sctp/ulpevent.c

@@ -425,6 +425,9 @@ struct sctp_ulpevent *sctp_ulpevent_make_send_failed(
 	struct sctp_send_failed *ssf;
 	struct sk_buff *skb;
 
+	/* Pull off any padding. */
+	int len = ntohs(chunk->chunk_hdr->length);
+
 	/* Make skb with more room so we can prepend notification.  */
 	skb = skb_copy_expand(chunk->skb,
 			      sizeof(struct sctp_send_failed), /* headroom */
@@ -434,7 +437,8 @@ struct sctp_ulpevent *sctp_ulpevent_make_send_failed(
 		goto fail;
 
 	/* Pull off the common chunk header and DATA header.  */
-	skb_pull(skb, sizeof(sctp_data_chunk_t));
+	skb_pull(skb, sizeof(struct sctp_data_chunk));
+	len -= sizeof(struct sctp_data_chunk);
 
 	/* Embed the event fields inside the cloned skb.  */
 	event = sctp_skb2event(skb);
@@ -475,7 +479,8 @@ struct sctp_ulpevent *sctp_ulpevent_make_send_failed(
 	 * This field is the total length of the notification data, including
 	 * the notification header.
 	 */
-	ssf->ssf_length = skb->len;
+	ssf->ssf_length = sizeof(struct sctp_send_failed) + len;
+	skb_trim(skb, ssf->ssf_length);
 
 	/* Socket Extensions for SCTP
 	 * 5.3.1.4 SCTP_SEND_FAILED
@@ -496,6 +501,11 @@ struct sctp_ulpevent *sctp_ulpevent_make_send_failed(
 	 */
 	memcpy(&ssf->ssf_info, &chunk->sinfo, sizeof(struct sctp_sndrcvinfo));
 
+	/* Per TSVWG discussion with Randy. Allow the application to
+	 * ressemble a fragmented message. 
+	 */
+	ssf->ssf_info.sinfo_flags = chunk->chunk_hdr->flags;
+
 	/* Socket Extensions for SCTP
 	 * 5.3.1.4 SCTP_SEND_FAILED
 	 *