[SCTP]: Force update of the rto when processing HB-ACK

When processing a HEARTBEAT-ACK it's possible that the transport rto timers will not be updated because a prior T3-RTX processing would have cleared the rto_pending flag on the transport. However, if we received a valid HEARTBEAT-ACK, we want to force update the rto variables, so re-set the rto_pending flag before calling sctp_transport_update_rto(). Signed-off-by: Vlad Yasevich <[email protected]> Signed-off-by: Sridhar Samudrala <[email protected]> Signed-off-by: David S. Miller <[email protected]>
cbmeeks · Jan 30, 2007 · e533ca1 · e533ca1
1 parent 2955652
commit e533ca1
Showing 1 changed file with 6 additions and 0 deletions.
diff --git a/net/sctp/sm_sideeffect.c b/net/sctp/sm_sideeffect.c
@@ -621,7 +621,13 @@ static void sctp_cmd_transport_on(sctp_cmd_seq_t *cmds,
 	/* The receiver of the HEARTBEAT ACK should also perform an
 	 * RTT measurement for that destination transport address
 	 * using the time value carried in the HEARTBEAT ACK chunk.
+	 * If the transport's rto_pending variable has been cleared,
+	 * it was most likely due to a retransmit.  However, we want
+	 * to re-enable it to properly update the rto.
 	 */
+	if (t->rto_pending == 0)
+		t->rto_pending = 1;
+
 	hbinfo = (sctp_sender_hb_info_t *) chunk->skb->data;
 	sctp_transport_update_rto(t, (jiffies - hbinfo->sent_at));