tcp: fix cc recovery re-entry and persist timer pop
Change-Id: I89e8052f2d2c36dd3de5255c4ee570722dc58227
Signed-off-by: Florin Coras <fcoras@cisco.com>
diff --git a/src/vnet/session/session.c b/src/vnet/session/session.c
index 6378fe8..9790ec2 100644
--- a/src/vnet/session/session.c
+++ b/src/vnet/session/session.c
@@ -491,9 +491,9 @@
application_t *app;
app = application_get_if_valid (s->app_index);
- if (PREDICT_FALSE (app == 0))
+ if (PREDICT_FALSE (!app))
{
- clib_warning ("invalid s->app_index = %d", s->app_index);
+ TCP_DBG ("invalid s->app_index = %d", s->app_index);
return 0;
}
diff --git a/src/vnet/session/session.h b/src/vnet/session/session.h
index 5e94c41..63ed83d 100644
--- a/src/vnet/session/session.h
+++ b/src/vnet/session/session.h
@@ -571,6 +571,15 @@
const transport_proto_vft_t * vft, u8 is_ip4,
u32 output_node);
+always_inline void
+transport_add_tx_event (transport_connection_t * tc)
+{
+ stream_session_t *s = session_get (tc->s_index, tc->thread_index);
+ if (svm_fifo_has_event (s->server_tx_fifo))
+ return;
+ session_send_io_evt_to_thread (s->server_tx_fifo, FIFO_EVENT_APP_TX);
+}
+
clib_error_t *vnet_session_enable_disable (vlib_main_t * vm, u8 is_en);
always_inline svm_msg_q_t *
diff --git a/src/vnet/tcp/tcp_input.c b/src/vnet/tcp/tcp_input.c
index 1e6a8eb..3f36b69 100644
--- a/src/vnet/tcp/tcp_input.c
+++ b/src/vnet/tcp/tcp_input.c
@@ -453,7 +453,11 @@
/* Karn's rule, part 1. Don't use retransmitted segments to estimate
* RTT because they're ambiguous. */
if (tcp_in_cong_recovery (tc) || tc->sack_sb.sacked_bytes)
- goto done;
+ {
+ if (tcp_in_recovery (tc))
+ return 0;
+ goto done;
+ }
if (tc->rtt_ts && seq_geq (ack, tc->rtt_seq))
{
@@ -479,7 +483,7 @@
tc->rtt_ts = 0;
/* If we got here something must've been ACKed so make sure boff is 0,
- * even if mrrt is not valid since we update the rto lower */
+ * even if mrtt is not valid since we update the rto lower */
tc->rto_boff = 0;
tcp_update_rto (tc);
@@ -1240,7 +1244,10 @@
* Legitimate ACK. 1) See if we can exit recovery
*/
/* XXX limit this only to first partial ack? */
- tcp_retransmit_timer_update (tc);
+ if (seq_lt (tc->snd_una, tc->snd_congestion))
+ tcp_retransmit_timer_force_update (tc);
+ else
+ tcp_retransmit_timer_update (tc);
if (seq_geq (tc->snd_una, tc->snd_congestion))
{
@@ -1272,6 +1279,7 @@
{
tc->cc_algo->rcv_ack (tc);
tc->tsecr_last_ack = tc->rcv_opts.tsecr;
+ transport_add_tx_event (&tc->connection);
return;
}
@@ -1322,10 +1330,11 @@
/* When we entered recovery, we reset snd_nxt to snd_una. Seems peer
* still has the data so accept the ack */
if (tcp_in_recovery (tc)
- && seq_leq (vnet_buffer (b)->tcp.ack_number, tc->snd_congestion)
- && seq_geq (vnet_buffer (b)->tcp.ack_number, tc->snd_una))
+ && seq_leq (vnet_buffer (b)->tcp.ack_number, tc->snd_congestion))
{
- tc->snd_una_max = tc->snd_nxt = vnet_buffer (b)->tcp.ack_number;
+ tc->snd_nxt = vnet_buffer (b)->tcp.ack_number;
+ if (seq_gt (tc->snd_nxt, tc->snd_una_max))
+ tc->snd_una_max = tc->snd_nxt;
goto process_ack;
}
diff --git a/src/vnet/tcp/tcp_output.c b/src/vnet/tcp/tcp_output.c
index b3c39b1..f37958e 100644
--- a/src/vnet/tcp/tcp_output.c
+++ b/src/vnet/tcp/tcp_output.c
@@ -1187,7 +1187,8 @@
tcp_push_hdr_i (tc, b, TCP_STATE_ESTABLISHED, /* compute opts */ 0,
/* burst */ 1);
tc->snd_una_max = tc->snd_nxt;
- ASSERT (seq_leq (tc->snd_una_max, tc->snd_una + tc->snd_wnd));
+ ASSERT (seq_leq (tc->snd_una_max, tc->snd_una + tc->snd_wnd
+ + tcp_fastrecovery_sent_1_smss (tc) * tc->snd_mss));
tcp_validate_txf_size (tc, tc->snd_una_max - tc->snd_una);
/* If not tracking an ACK, start tracking */
if (tc->rtt_ts == 0 && !tcp_in_cong_recovery (tc))
@@ -1451,15 +1452,19 @@
return;
}
- /* Shouldn't be here */
+ /* Shouldn't be here. This condition is tricky because it has to take
+ * into account boff > 0 due to persist timeout. */
if ((tc->rto_boff == 0 && tc->snd_una == tc->snd_una_max)
- || (tc->rto_boff > 0 && seq_geq (tc->snd_una, tc->snd_congestion)))
+ || (tc->rto_boff > 0 && seq_geq (tc->snd_una, tc->snd_congestion)
+ && !tcp_flight_size (tc)))
{
- tcp_recovery_off (tc);
+ ASSERT (!tcp_in_recovery (tc));
+ tc->rto_boff = 0;
return;
}
- /* We're not in recovery so make sure rto_boff is 0 */
+ /* We're not in recovery so make sure rto_boff is 0. Can be non 0 due
+ * to persist timer timeout */
if (!tcp_in_recovery (tc) && tc->rto_boff > 0)
{
tc->rto_boff = 0;
@@ -1474,10 +1479,15 @@
if (tc->rto_boff == 1)
tcp_rxt_timeout_cc (tc);
+ /* If we've sent beyond snd_congestion, update it */
+ if (seq_gt (tc->snd_una_max, tc->snd_congestion))
+ tc->snd_congestion = tc->snd_una_max;
+
tc->snd_una_max = tc->snd_nxt = tc->snd_una;
tc->rto = clib_min (tc->rto << 1, TCP_RTO_MAX);
- /* Send one segment. Note that n_bytes may be zero due to buffer shortfall */
+ /* Send one segment. Note that n_bytes may be zero due to buffer
+ * shortfall */
n_bytes = tcp_prepare_retransmit_segment (tc, 0, tc->snd_mss, &b);
/* TODO be less aggressive about this */
@@ -1485,7 +1495,7 @@
if (n_bytes == 0)
{
- tcp_retransmit_timer_set (tc);
+ tcp_retransmit_timer_force_update (tc);
return;
}
@@ -1496,7 +1506,7 @@
tc->snd_rxt_ts = tcp_time_now ();
tcp_enqueue_to_output (vm, b, bi, tc->c_is_ip4);
- tcp_retransmit_timer_update (tc);
+ tcp_retransmit_timer_force_update (tc);
}
/* Retransmit for SYN */
else if (tc->state == TCP_STATE_SYN_SENT)
@@ -1632,7 +1642,10 @@
* Try to force the first unsent segment (or buffer)
*/
if (PREDICT_FALSE (tcp_get_free_buffer_index (tm, &bi)))
- return;
+ {
+ tcp_persist_timer_set (tc);
+ return;
+ }
b = vlib_get_buffer (vm, bi);
data = tcp_init_buffer (vm, b);