sctp: Correctly implement Fast Recovery cwnd manipulations.

Correctly keep track of Fast Recovery state and do not reduce
congestion window multiple times during sucht state.

Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Tested-by: Wei Yongjun <yjwei@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

2 files changed, 39 insertions, 13 deletions

diff --git a/include/net/sctp/structs.h b/include/net/sctp/structs.h
index 917d425f0542..67592072a32e 100644
--- a/include/net/sctp/structs.h
+++ b/include/net/sctp/structs.h
@@ -902,7 +902,10 @@ struct sctp_transport {
          *              calculation completes (i.e. the DATA chunk
          *              is SACK'd) clear this flag.
          */
-        int rto_pending;
+        __u8 rto_pending;
+
+        /* Flag to track the current fast recovery state */
+        __u8 fast_recovery;
 
         /*
          * These are the congestion stats.
@@ -921,6 +924,9 @@ struct sctp_transport {
         /* Data that has been sent, but not acknowledged. */
         __u32 flight_size;
 
+        /* TSN marking the fast recovery exit point */
+        __u32 fast_recovery_exit;
+
         /* Destination */
         struct dst_entry *dst;
         /* Source address. */
diff --git a/net/sctp/transport.c b/net/sctp/transport.c
index 62082e7b7972..9647fb277221 100644
--- a/net/sctp/transport.c
+++ b/net/sctp/transport.c
@@ -79,6 +79,7 @@ static struct sctp_transport *sctp_transport_init(struct sctp_transport *peer,
         peer->rttvar = 0;
         peer->srtt = 0;
         peer->rto_pending = 0;
+        peer->fast_recovery = 0;
 
         peer->last_time_heard = jiffies;
         peer->last_time_used = jiffies;
@@ -403,11 +404,16 @@ void sctp_transport_raise_cwnd(struct sctp_transport *transport,
         cwnd = transport->cwnd;
         flight_size = transport->flight_size;
 
+        /* See if we need to exit Fast Recovery first */
+        if (transport->fast_recovery &&
+            TSN_lte(transport->fast_recovery_exit, sack_ctsn))
+                transport->fast_recovery = 0;
+
         /* The appropriate cwnd increase algorithm is performed if, and only
-         * if the cumulative TSN has advanced and the congestion window is
+         * if the cumulative TSN whould advanced and the congestion window is
          * being fully utilized.
          */
-        if ((transport->asoc->ctsn_ack_point >= sack_ctsn) ||
+        if (TSN_lte(sack_ctsn, transport->asoc->ctsn_ack_point) ||
             (flight_size < cwnd))
                 return;
 
@@ -416,17 +422,23 @@ void sctp_transport_raise_cwnd(struct sctp_transport *transport,
         pmtu = transport->asoc->pathmtu;
 
         if (cwnd <= ssthresh) {
-                /* RFC 2960 7.2.1, sctpimpguide-05 2.14.2 When cwnd is less
-                 * than or equal to ssthresh an SCTP endpoint MUST use the
-                 * slow start algorithm to increase cwnd only if the current
-                 * congestion window is being fully utilized and an incoming
-                 * SACK advances the Cumulative TSN Ack Point. Only when these
-                 * two conditions are met can the cwnd be increased otherwise
-                 * the cwnd MUST not be increased. If these conditions are met
-                 * then cwnd MUST be increased by at most the lesser of
-                 * 1) the total size of the previously outstanding DATA
-                 * chunk(s) acknowledged, and 2) the destination's path MTU.
+                /* RFC 4960 7.2.1
+                 * o  When cwnd is less than or equal to ssthresh, an SCTP
+                 *    endpoint MUST use the slow-start algorithm to increase
+                 *    cwnd only if the current congestion window is being fully
+                 *    utilized, an incoming SACK advances the Cumulative TSN
+                 *    Ack Point, and the data sender is not in Fast Recovery.
+                 *    Only when these three conditions are met can the cwnd be
+                 *    increased; otherwise, the cwnd MUST not be increased.
+                 *    If these conditions are met, then cwnd MUST be increased
+                 *    by, at most, the lesser of 1) the total size of the
+                 *    previously outstanding DATA chunk(s) acknowledged, and
+                 *    2) the destination's path MTU.  This upper bound protects
+                 *    against the ACK-Splitting attack outlined in [SAVAGE99].
                  */
+                if (transport->fast_recovery)
+                        return;
+
                 if (bytes_acked > pmtu)
                         cwnd += pmtu;
                 else
@@ -502,6 +514,13 @@ void sctp_transport_lower_cwnd(struct sctp_transport *transport,
                  *      cwnd = ssthresh
                  *      partial_bytes_acked = 0
                  */
+                if (transport->fast_recovery)
+                        return;
+
+                /* Mark Fast recovery */
+                transport->fast_recovery = 1;
+                transport->fast_recovery_exit = transport->asoc->next_tsn - 1;
+
                 transport->ssthresh = max(transport->cwnd/2,
                                           4*transport->asoc->pathmtu);
                 transport->cwnd = transport->ssthresh;
@@ -586,6 +605,7 @@ void sctp_transport_reset(struct sctp_transport *t)
         t->flight_size = 0;
         t->error_count = 0;
         t->rto_pending = 0;
+        t->fast_recovery = 0;
 
         /* Initialize the state information for SFR-CACC */
         t->cacc.changeover_active = 0;