BUG_ON(len > skb->len);
- tcp_clear_retrans_hints_partial(tp);
+ tcp_clear_all_retrans_hints(tp);
nsize = skb_headlen(skb) - len;
if (nsize < 0)
nsize = 0;
tp->packets_out -= tcp_skb_pcount(next_skb);
/* changed transmit queue under us so clear hints */
- tcp_clear_retrans_hints_partial(tp);
+ tcp_clear_all_retrans_hints(tp);
sk_wmem_free_skb(sk, next_skb);
}
struct tcp_sock *tp = tcp_sk(sk);
struct sk_buff *skb;
unsigned int mss = tcp_current_mss(sk, 0);
- int lost = 0;
+ u32 prior_lost = tp->lost_out;
tcp_for_write_queue(skb, sk) {
if (skb == tcp_send_head(sk))
TCP_SKB_CB(skb)->sacked &= ~TCPCB_SACKED_RETRANS;
tp->retrans_out -= tcp_skb_pcount(skb);
}
- if (!(TCP_SKB_CB(skb)->sacked & TCPCB_LOST)) {
- TCP_SKB_CB(skb)->sacked |= TCPCB_LOST;
- tp->lost_out += tcp_skb_pcount(skb);
- lost = 1;
- }
+ tcp_skb_mark_lost_uncond_verify(tp, skb);
}
}
tcp_clear_all_retrans_hints(tp);
- if (!lost)
+ if (prior_lost == tp->lost_out)
return;
if (tcp_is_reno(tp))
return err;
}
-/* This gets called after a retransmit timeout, and the initially
- * retransmitted data is acknowledged. It tries to continue
- * resending the rest of the retransmit queue, until either
- * we've sent it all or the congestion window limit is reached.
- * If doing SACK, the first ACK which comes back for a timeout
- * based retransmit packet might feed us FACK information again.
- * If so, we use it to avoid unnecessarily retransmissions.
- */
-void tcp_xmit_retransmit_queue(struct sock *sk)
+static int tcp_can_forward_retransmit(struct sock *sk)
{
const struct inet_connection_sock *icsk = inet_csk(sk);
struct tcp_sock *tp = tcp_sk(sk);
- struct sk_buff *skb;
- int packet_cnt;
-
- if (tp->retransmit_skb_hint) {
- skb = tp->retransmit_skb_hint;
- packet_cnt = tp->retransmit_cnt_hint;
- } else {
- skb = tcp_write_queue_head(sk);
- packet_cnt = 0;
- }
-
- /* First pass: retransmit lost packets. */
- if (tp->lost_out) {
- tcp_for_write_queue_from(skb, sk) {
- __u8 sacked = TCP_SKB_CB(skb)->sacked;
-
- if (skb == tcp_send_head(sk))
- break;
- /* we could do better than to assign each time */
- tp->retransmit_skb_hint = skb;
- tp->retransmit_cnt_hint = packet_cnt;
-
- /* Assume this retransmit will generate
- * only one packet for congestion window
- * calculation purposes. This works because
- * tcp_retransmit_skb() will chop up the
- * packet to be MSS sized and all the
- * packet counting works out.
- */
- if (tcp_packets_in_flight(tp) >= tp->snd_cwnd)
- return;
-
- if (sacked & TCPCB_LOST) {
- if (!(sacked & (TCPCB_SACKED_ACKED|TCPCB_SACKED_RETRANS))) {
- int mib_idx;
-
- if (tcp_retransmit_skb(sk, skb)) {
- tp->retransmit_skb_hint = NULL;
- return;
- }
- if (icsk->icsk_ca_state != TCP_CA_Loss)
- mib_idx = LINUX_MIB_TCPFASTRETRANS;
- else
- mib_idx = LINUX_MIB_TCPSLOWSTARTRETRANS;
- NET_INC_STATS_BH(sock_net(sk), mib_idx);
-
- if (skb == tcp_write_queue_head(sk))
- inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS,
- inet_csk(sk)->icsk_rto,
- TCP_RTO_MAX);
- }
-
- packet_cnt += tcp_skb_pcount(skb);
- if (packet_cnt >= tp->lost_out)
- break;
- }
- }
- }
-
- /* OK, demanded retransmission is finished. */
/* Forward retransmissions are possible only during Recovery. */
if (icsk->icsk_ca_state != TCP_CA_Recovery)
- return;
+ return 0;
/* No forward retransmissions in Reno are possible. */
if (tcp_is_reno(tp))
- return;
+ return 0;
/* Yeah, we have to make difficult choice between forward transmission
* and retransmission... Both ways have their merits...
*/
if (tcp_may_send_now(sk))
- return;
+ return 0;
- /* If nothing is SACKed, highest_sack in the loop won't be valid */
- if (!tp->sacked_out)
- return;
+ return 1;
+}
- if (tp->forward_skb_hint)
- skb = tp->forward_skb_hint;
+/* This gets called after a retransmit timeout, and the initially
+ * retransmitted data is acknowledged. It tries to continue
+ * resending the rest of the retransmit queue, until either
+ * we've sent it all or the congestion window limit is reached.
+ * If doing SACK, the first ACK which comes back for a timeout
+ * based retransmit packet might feed us FACK information again.
+ * If so, we use it to avoid unnecessarily retransmissions.
+ */
+void tcp_xmit_retransmit_queue(struct sock *sk)
+{
+ const struct inet_connection_sock *icsk = inet_csk(sk);
+ struct tcp_sock *tp = tcp_sk(sk);
+ struct sk_buff *skb;
+ struct sk_buff *hole = NULL;
+ int mib_idx;
+ int fwd_rexmitting = 0;
+
+ if (!tp->lost_out)
+ tp->retransmit_high = tp->snd_una;
+
+ if (tp->retransmit_skb_hint)
+ skb = tp->retransmit_skb_hint;
else
skb = tcp_write_queue_head(sk);
+ /* First pass: retransmit lost packets. */
tcp_for_write_queue_from(skb, sk) {
- if (skb == tcp_send_head(sk))
- break;
- tp->forward_skb_hint = skb;
+ __u8 sacked = TCP_SKB_CB(skb)->sacked;
- if (!before(TCP_SKB_CB(skb)->seq, tcp_highest_sack_seq(tp)))
+ if (skb == tcp_send_head(sk))
break;
+ /* we could do better than to assign each time */
+ if (hole == NULL)
+ tp->retransmit_skb_hint = skb;
+ /* Assume this retransmit will generate
+ * only one packet for congestion window
+ * calculation purposes. This works because
+ * tcp_retransmit_skb() will chop up the
+ * packet to be MSS sized and all the
+ * packet counting works out.
+ */
if (tcp_packets_in_flight(tp) >= tp->snd_cwnd)
- break;
+ return;
+
+ if (fwd_rexmitting) {
+begin_fwd:
+ if (!before(TCP_SKB_CB(skb)->seq, tcp_highest_sack_seq(tp)))
+ break;
+ mib_idx = LINUX_MIB_TCPFORWARDRETRANS;
+
+ } else if (!before(TCP_SKB_CB(skb)->seq, tp->retransmit_high)) {
+ if (!tcp_can_forward_retransmit(sk))
+ break;
+ /* Backtrack if necessary to non-L'ed skb */
+ if (hole != NULL) {
+ skb = hole;
+ hole = NULL;
+ }
+ fwd_rexmitting = 1;
+ goto begin_fwd;
- if (TCP_SKB_CB(skb)->sacked & TCPCB_TAGBITS)
+ } else if (!(sacked & TCPCB_LOST)) {
+ if (hole == NULL && !(sacked & TCPCB_SACKED_RETRANS))
+ hole = skb;
continue;
- /* Ok, retransmit it. */
- if (tcp_retransmit_skb(sk, skb)) {
- tp->forward_skb_hint = NULL;
- break;
+ } else {
+ if (icsk->icsk_ca_state != TCP_CA_Loss)
+ mib_idx = LINUX_MIB_TCPFASTRETRANS;
+ else
+ mib_idx = LINUX_MIB_TCPSLOWSTARTRETRANS;
}
+ if (sacked & (TCPCB_SACKED_ACKED|TCPCB_SACKED_RETRANS))
+ continue;
+
+ if (tcp_retransmit_skb(sk, skb))
+ return;
+ NET_INC_STATS_BH(sock_net(sk), mib_idx);
+
if (skb == tcp_write_queue_head(sk))
inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS,
inet_csk(sk)->icsk_rto,
TCP_RTO_MAX);
-
- NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPFORWARDRETRANS);
}
}