#define FLAG_FORWARD_PROGRESS (FLAG_ACKED|FLAG_DATA_SACKED)
#define FLAG_ANY_PROGRESS (FLAG_FORWARD_PROGRESS|FLAG_SND_UNA_ADVANCED)
-#define IsReno(tp) ((tp)->rx_opt.sack_ok == 0)
-#define IsFack(tp) ((tp)->rx_opt.sack_ok & 2)
-#define IsDSack(tp) ((tp)->rx_opt.sack_ok & 4)
-
#define IsSackFrto() (sysctl_tcp_frto == 0x2)
#define TCP_REMNANT (TCP_FLAG_FIN|TCP_FLAG_URG|TCP_FLAG_SYN|TCP_FLAG_PSH)
}
}
+/*
+ * Packet counting of FACK is based on in-order assumptions, therefore TCP
+ * disables it when reordering is detected
+ */
+static void tcp_disable_fack(struct tcp_sock *tp)
+{
+ tp->rx_opt.sack_ok &= ~2;
+}
+
+/* Take a notice that peer is sending DSACKs */
+static void tcp_dsack_seen(struct tcp_sock *tp)
+{
+ tp->rx_opt.sack_ok |= 4;
+}
+
/* Initialize metrics on socket. */
static void tcp_init_metrics(struct sock *sk)
}
if (dst_metric(dst, RTAX_REORDERING) &&
tp->reordering != dst_metric(dst, RTAX_REORDERING)) {
- tp->rx_opt.sack_ok &= ~2;
+ tcp_disable_fack(tp);
tp->reordering = dst_metric(dst, RTAX_REORDERING);
}
/* This exciting event is worth to be remembered. 8) */
if (ts)
NET_INC_STATS_BH(LINUX_MIB_TCPTSREORDER);
- else if (IsReno(tp))
+ else if (tcp_is_reno(tp))
NET_INC_STATS_BH(LINUX_MIB_TCPRENOREORDER);
- else if (IsFack(tp))
+ else if (tcp_is_fack(tp))
NET_INC_STATS_BH(LINUX_MIB_TCPFACKREORDER);
else
NET_INC_STATS_BH(LINUX_MIB_TCPSACKREORDER);
tp->sacked_out,
tp->undo_marker ? tp->undo_retrans : 0);
#endif
- /* Disable FACK yet. */
- tp->rx_opt.sack_ok &= ~2;
+ tcp_disable_fack(tp);
}
}
* for retransmitted and already SACKed segment -> reordering..
* Both of these heuristics are not used in Loss state, when we cannot
* account for retransmits accurately.
+ *
+ * SACK block validation.
+ * ----------------------
+ *
+ * SACK block range validation checks that the received SACK block fits to
+ * the expected sequence limits, i.e., it is between SND.UNA and SND.NXT.
+ * Note that SND.UNA is not included to the range though being valid because
+ * it means that the receiver is rather inconsistent with itself (reports
+ * SACK reneging when it should advance SND.UNA).
+ *
+ * Implements also blockage to start_seq wrap-around. Problem lies in the
+ * fact that though start_seq (s) is before end_seq (i.e., not reversed),
+ * there's no guarantee that it will be before snd_nxt (n). The problem
+ * happens when start_seq resides between end_seq wrap (e_w) and snd_nxt
+ * wrap (s_w):
+ *
+ * <- outs wnd -> <- wrapzone ->
+ * u e n u_w e_w s n_w
+ * | | | | | | |
+ * |<------------+------+----- TCP seqno space --------------+---------->|
+ * ...-- <2^31 ->| |<--------...
+ * ...---- >2^31 ------>| |<--------...
+ *
+ * Current code wouldn't be vulnerable but it's better still to discard such
+ * crazy SACK blocks. Doing this check for start_seq alone closes somewhat
+ * similar case (end_seq after snd_nxt wrap) as earlier reversed check in
+ * snd_nxt wrap -> snd_una region will then become "well defined", i.e.,
+ * equal to the ideal case (infinite seqno space without wrap caused issues).
+ *
+ * With D-SACK the lower bound is extended to cover sequence space below
+ * SND.UNA down to undo_marker, which is the last point of interest. Yet
+ * again, DSACK block must not to go across snd_una (for the same reason as
+ * for the normal SACK blocks, explained above). But there all simplicity
+ * ends, TCP might receive valid D-SACKs below that. As long as they reside
+ * fully below undo_marker they do not affect behavior in anyway and can
+ * therefore be safely ignored. In rare cases (which are more or less
+ * theoretical ones), the D-SACK will nicely cross that boundary due to skb
+ * fragmentation and packet reordering past skb's retransmission. To consider
+ * them correctly, the acceptable range must be extended even more though
+ * the exact amount is rather hard to quantify. However, tp->max_window can
+ * be used as an exaggerated estimate.
*/
+static int tcp_is_sackblock_valid(struct tcp_sock *tp, int is_dsack,
+ u32 start_seq, u32 end_seq)
+{
+ /* Too far in future, or reversed (interpretation is ambiguous) */
+ if (after(end_seq, tp->snd_nxt) || !before(start_seq, end_seq))
+ return 0;
+
+ /* Nasty start_seq wrap-around check (see comments above) */
+ if (!before(start_seq, tp->snd_nxt))
+ return 0;
+
+ /* In outstanding window? ...This is valid exit for DSACKs too.
+ * start_seq == snd_una is non-sensical (see comments above)
+ */
+ if (after(start_seq, tp->snd_una))
+ return 1;
+
+ if (!is_dsack || !tp->undo_marker)
+ return 0;
+
+ /* ...Then it's D-SACK, and must reside below snd_una completely */
+ if (!after(end_seq, tp->snd_una))
+ return 0;
+
+ if (!before(start_seq, tp->undo_marker))
+ return 1;
+
+ /* Too old */
+ if (!after(end_seq, tp->undo_marker))
+ return 0;
+
+ /* Undo_marker boundary crossing (overestimates a lot). Known already:
+ * start_seq < undo_marker and end_seq >= undo_marker.
+ */
+ return !before(start_seq, end_seq - tp->max_window);
+}
+
+
static int tcp_check_dsack(struct tcp_sock *tp, struct sk_buff *ack_skb,
struct tcp_sack_block_wire *sp, int num_sacks,
u32 prior_snd_una)
if (before(start_seq_0, TCP_SKB_CB(ack_skb)->ack_seq)) {
dup_sack = 1;
- tp->rx_opt.sack_ok |= 4;
+ tcp_dsack_seen(tp);
NET_INC_STATS_BH(LINUX_MIB_TCPDSACKRECV);
} else if (num_sacks > 1) {
u32 end_seq_1 = ntohl(get_unaligned(&sp[1].end_seq));
if (!after(end_seq_0, end_seq_1) &&
!before(start_seq_0, start_seq_1)) {
dup_sack = 1;
- tp->rx_opt.sack_ok |= 4;
+ tcp_dsack_seen(tp);
NET_INC_STATS_BH(LINUX_MIB_TCPDSACKOFORECV);
}
}
int fack_count;
int dup_sack = (found_dup_sack && (i == first_sack_index));
+ if (!tcp_is_sackblock_valid(tp, dup_sack, start_seq, end_seq)) {
+ if (dup_sack) {
+ if (!tp->undo_marker)
+ NET_INC_STATS_BH(LINUX_MIB_TCPDSACKIGNOREDNOUNDO);
+ else
+ NET_INC_STATS_BH(LINUX_MIB_TCPDSACKIGNOREDOLD);
+ } else
+ NET_INC_STATS_BH(LINUX_MIB_TCPSACKDISCARD);
+ continue;
+ }
+
skb = cached_skb;
fack_count = cached_fack_count;
continue;
if ((TCP_SKB_CB(skb)->sacked&TCPCB_SACKED_RETRANS) &&
after(lost_retrans, TCP_SKB_CB(skb)->ack_seq) &&
- (IsFack(tp) ||
+ (tcp_is_fack(tp) ||
!before(lost_retrans,
TCP_SKB_CB(skb)->ack_seq + tp->reordering *
tp->mss_cache))) {
tp->lost_out = 0;
tp->retrans_out = 0;
- if (IsReno(tp))
+ if (tcp_is_reno(tp))
tcp_reset_reno_sack(tp);
tcp_for_write_queue(skb, sk) {
static inline int tcp_fackets_out(struct tcp_sock *tp)
{
- return IsReno(tp) ? tp->sacked_out+1 : tp->fackets_out;
+ return tcp_is_reno(tp) ? tp->sacked_out+1 : tp->fackets_out;
}
static inline int tcp_skb_timedout(struct sock *sk, struct sk_buff *skb)
{
struct tcp_sock *tp = tcp_sk(sk);
- if (IsFack(tp)) {
+ if (tcp_is_fack(tp)) {
int lost = tp->fackets_out - tp->reordering;
if (lost <= 0)
lost = 1;
* Hence, we can detect timed out packets during fast
* retransmit without falling to slow start.
*/
- if (!IsReno(tp) && tcp_head_timedout(sk)) {
+ if (!tcp_is_reno(tp) && tcp_head_timedout(sk)) {
struct sk_buff *skb;
skb = tp->scoreboard_skb_hint ? tp->scoreboard_skb_hint
int decr = tp->snd_cwnd_cnt + 1;
if ((flag&(FLAG_ANY_PROGRESS|FLAG_DSACKING_ACK)) ||
- (IsReno(tp) && !(flag&FLAG_NOT_DUP))) {
+ (tcp_is_reno(tp) && !(flag&FLAG_NOT_DUP))) {
tp->snd_cwnd_cnt = decr&1;
decr >>= 1;
NET_INC_STATS_BH(LINUX_MIB_TCPFULLUNDO);
tp->undo_marker = 0;
}
- if (tp->snd_una == tp->high_seq && IsReno(tp)) {
+ if (tp->snd_una == tp->high_seq && tcp_is_reno(tp)) {
/* Hold old state until something *above* high_seq
* is ACKed. For Reno it is MUST to prevent false
* fast retransmits (RFC2582). SACK TCP is safe. */
{
struct tcp_sock *tp = tcp_sk(sk);
/* Partial ACK arrived. Force Hoe's retransmit. */
- int failed = IsReno(tp) || tp->fackets_out>tp->reordering;
+ int failed = tcp_is_reno(tp) || tp->fackets_out>tp->reordering;
if (tcp_may_undo(tp)) {
/* Plain luck! Hole if filled with delayed
NET_INC_STATS_BH(LINUX_MIB_TCPLOSSUNDO);
inet_csk(sk)->icsk_retransmits = 0;
tp->undo_marker = 0;
- if (!IsReno(tp))
+ if (tcp_is_sack(tp))
tcp_set_ca_state(sk, TCP_CA_Open);
return 1;
}
if (inet_csk(sk)->icsk_ca_state != TCP_CA_CWR) {
int state = TCP_CA_Open;
- if (tp->sacked_out || tp->retrans_out || tp->undo_marker)
+ if (tcp_left_out(tp) || tp->retrans_out || tp->undo_marker)
state = TCP_CA_Disorder;
if (inet_csk(sk)->icsk_ca_state != state) {
* tcp_xmit_retransmit_queue().
*/
static void
-tcp_fastretrans_alert(struct sock *sk, int prior_packets, int flag)
+tcp_fastretrans_alert(struct sock *sk, int pkts_acked, int flag)
{
struct inet_connection_sock *icsk = inet_csk(sk);
struct tcp_sock *tp = tcp_sk(sk);
if (!tp->undo_marker ||
/* For SACK case do not Open to allow to undo
* catching for all duplicate ACKs. */
- IsReno(tp) || tp->snd_una != tp->high_seq) {
+ tcp_is_reno(tp) || tp->snd_una != tp->high_seq) {
tp->undo_marker = 0;
tcp_set_ca_state(sk, TCP_CA_Open);
}
break;
case TCP_CA_Recovery:
- if (IsReno(tp))
+ if (tcp_is_reno(tp))
tcp_reset_reno_sack(tp);
if (tcp_try_undo_recovery(sk))
return;
switch (icsk->icsk_ca_state) {
case TCP_CA_Recovery:
if (!(flag & FLAG_SND_UNA_ADVANCED)) {
- if (IsReno(tp) && is_dupack)
+ if (tcp_is_reno(tp) && is_dupack)
tcp_add_reno_sack(sk);
- } else {
- int acked = prior_packets - tp->packets_out;
- if (IsReno(tp))
- tcp_remove_reno_sacks(sk, acked);
- do_lost = tcp_try_undo_partial(sk, acked);
- }
+ } else
+ do_lost = tcp_try_undo_partial(sk, pkts_acked);
break;
case TCP_CA_Loss:
if (flag&FLAG_DATA_ACKED)
return;
/* Loss is undone; fall through to processing in Open state. */
default:
- if (IsReno(tp)) {
+ if (tcp_is_reno(tp)) {
if (flag & FLAG_SND_UNA_ADVANCED)
tcp_reset_reno_sack(tp);
if (is_dupack)
/* Otherwise enter Recovery state */
- if (IsReno(tp))
+ if (tcp_is_reno(tp))
NET_INC_STATS_BH(LINUX_MIB_TCPRENORECOVERY);
else
NET_INC_STATS_BH(LINUX_MIB_TCPSACKRECOVERY);
/* Restart timer after forward progress on connection.
* RFC2988 recommends to restart timer to now+rto.
*/
-
-static void tcp_ack_packets_out(struct sock *sk)
+static void tcp_rearm_rto(struct sock *sk)
{
struct tcp_sock *tp = tcp_sk(sk);
last_ackt = skb->tstamp;
}
tcp_dec_pcount_approx(&tp->fackets_out, skb);
- tcp_packets_out_dec(tp, skb);
+ tp->packets_out -= tcp_skb_pcount(skb);
tcp_unlink_write_queue(skb, sk);
sk_stream_free_skb(sk, skb);
clear_all_retrans_hints(tp);
= inet_csk(sk)->icsk_ca_ops;
tcp_ack_update_rtt(sk, acked, seq_rtt);
- tcp_ack_packets_out(sk);
+ tcp_rearm_rto(sk);
+
+ if (tcp_is_reno(tp))
+ tcp_remove_reno_sacks(sk, pkts_acked);
if (ca_ops->pkts_acked) {
s32 rtt_us = -1;
BUG_TRAP((int)tp->sacked_out >= 0);
BUG_TRAP((int)tp->lost_out >= 0);
BUG_TRAP((int)tp->retrans_out >= 0);
- if (!tp->packets_out && tp->rx_opt.sack_ok) {
+ if (!tp->packets_out && tcp_is_sack(tp)) {
const struct inet_connection_sock *icsk = inet_csk(sk);
if (tp->lost_out) {
printk(KERN_DEBUG "Leak l=%u %d\n",
return 1;
}
- if (!IsSackFrto() || IsReno(tp)) {
+ if (!IsSackFrto() || tcp_is_reno(tp)) {
/* RFC4138 shortcoming in step 2; should also have case c):
* ACK isn't duplicate nor advances window, e.g., opposite dir
* data, winupdate
if ((flag & FLAG_DATA_ACKED) && !frto_cwnd &&
tcp_may_raise_cwnd(sk, flag))
tcp_cong_avoid(sk, ack, prior_in_flight, 0);
- tcp_fastretrans_alert(sk, prior_packets, flag);
+ tcp_fastretrans_alert(sk, prior_packets - tp->packets_out, flag);
} else {
if ((flag & FLAG_DATA_ACKED) && !frto_cwnd)
tcp_cong_avoid(sk, ack, prior_in_flight, 1);
* Probably, we should reset in this case. For now drop them.
*/
__skb_queue_purge(&tp->out_of_order_queue);
- if (tp->rx_opt.sack_ok)
+ if (tcp_is_sack(tp))
tcp_sack_reset(&tp->rx_opt);
sk_stream_mem_reclaim(sk);
static void tcp_dsack_set(struct tcp_sock *tp, u32 seq, u32 end_seq)
{
- if (tp->rx_opt.sack_ok && sysctl_tcp_dsack) {
+ if (tcp_is_sack(tp) && sysctl_tcp_dsack) {
if (before(seq, tp->rcv_nxt))
NET_INC_STATS_BH(LINUX_MIB_TCPDSACKOLDSENT);
else
NET_INC_STATS_BH(LINUX_MIB_DELAYEDACKLOST);
tcp_enter_quickack_mode(sk);
- if (tp->rx_opt.sack_ok && sysctl_tcp_dsack) {
+ if (tcp_is_sack(tp) && sysctl_tcp_dsack) {
u32 end_seq = TCP_SKB_CB(skb)->end_seq;
if (after(TCP_SKB_CB(skb)->end_seq, tp->rcv_nxt))
if (!skb_peek(&tp->out_of_order_queue)) {
/* Initial out of order segment, build 1 SACK. */
- if (tp->rx_opt.sack_ok) {
+ if (tcp_is_sack(tp)) {
tp->rx_opt.num_sacks = 1;
tp->rx_opt.dsack = 0;
tp->rx_opt.eff_sacks = 1;
}
add_sack:
- if (tp->rx_opt.sack_ok)
+ if (tcp_is_sack(tp))
tcp_sack_new_ofo_skb(sk, seq, end_seq);
}
}
* is in a sad state like this, we care only about integrity
* of the connection not performance.
*/
- if (tp->rx_opt.sack_ok)
+ if (tcp_is_sack(tp))
tcp_sack_reset(&tp->rx_opt);
sk_stream_mem_reclaim(sk);
}
tp->tcp_header_len = sizeof(struct tcphdr);
}
- if (tp->rx_opt.sack_ok && sysctl_tcp_fack)
- tp->rx_opt.sack_ok |= 2;
+ if (tcp_is_sack(tp) && sysctl_tcp_fack)
+ tcp_enable_fack(tp);
tcp_mtup_init(sk);
tcp_sync_mss(sk, icsk->icsk_pmtu_cookie);