sctp: adjust cwnd increase in Congestion Avoidance phase

RFC4960 Errata 3.26 identified that at the same time RFC4960 states that
cwnd should never grow more than 1*MTU per RTT, Section 7.2.2 was
underspecified and as described could allow increasing cwnd more than
that.

This patch updates it so partial_bytes_acked is maxed to cwnd if
flight_size doesn't reach cwnd, protecting it from such case.

See-also: https://tools.ietf.org/html/draft-ietf-tsvwg-rfc4960-errata-01#section-3.26
Signed-off-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

diff --git a/net/sctp/transport.c b/net/sctp/transport.c
index 9d35894..e3ebf04 100644 (file)
--- a/net/sctp/transport.c
+++ b/net/sctp/transport.c
@@ -405,13 +405,6 @@ void sctp_transport_raise_cwnd(struct sctp_transport *transport,
            TSN_lte(asoc->fast_recovery_exit, sack_ctsn))
                asoc->fast_recovery = 0;
 
-       /* The appropriate cwnd increase algorithm is performed if, and only
-        * if the congestion window is being fully utilized.
-        * Note that RFC4960 Errata 3.22 removed the other condition.
-        */
-       if (flight_size < cwnd)
-               return;
-
        ssthresh = transport->ssthresh;
        pba = transport->partial_bytes_acked;
        pmtu = transport->asoc->pathmtu;
@@ -434,6 +427,14 @@ void sctp_transport_raise_cwnd(struct sctp_transport *transport,
                if (asoc->fast_recovery)
                        return;
 
+               /* The appropriate cwnd increase algorithm is performed
+                * if, and only if the congestion window is being fully
+                * utilized.  Note that RFC4960 Errata 3.22 removed the
+                * other condition on ctsn moving.
+                */
+               if (flight_size < cwnd)
+                       return;
+
                if (bytes_acked > pmtu)
                        cwnd += pmtu;
                else
@@ -451,6 +452,13 @@ void sctp_transport_raise_cwnd(struct sctp_transport *transport,
                 * acknowledged by the new Cumulative TSN Ack and by Gap
                 * Ack Blocks. (updated by RFC4960 Errata 3.22)
                 *
+                * When partial_bytes_acked is greater than cwnd and
+                * before the arrival of the SACK the sender had less
+                * bytes of data outstanding than cwnd (i.e., before
+                * arrival of the SACK, flightsize was less than cwnd),
+                * reset partial_bytes_acked to cwnd. (RFC 4960 Errata
+                * 3.26)
+                *
                 * When partial_bytes_acked is equal to or greater than
                 * cwnd and before the arrival of the SACK the sender
                 * had cwnd or more bytes of data outstanding (i.e.,
@@ -460,7 +468,9 @@ void sctp_transport_raise_cwnd(struct sctp_transport *transport,
                 * increased by MTU. (RFC 4960 Errata 3.12)
                 */
                pba += bytes_acked;
-               if (pba >= cwnd) {
+               if (pba > cwnd && flight_size < cwnd)
+                       pba = cwnd;
+               if (pba >= cwnd && flight_size >= cwnd) {
                        pba = pba - cwnd;
                        cwnd += pmtu;
                }