/* * INET An implementation of the TCP/IP protocol suite for the LINUX * operating system. INET is implemented using the BSD Socket * interface as the means of communication with the user level. * * Implementation of the Transmission Control Protocol(TCP). * * Version: $Id: tcp_input.c,v 1.243 2002/02/01 22:01:04 davem Exp $ * * Authors: Ross Biro * Fred N. van Kempen, * Mark Evans, * Corey Minyard * Florian La Roche, * Charles Hedrick, * Linus Torvalds, * Alan Cox, * Matthew Dillon, * Arnt Gulbrandsen, * Jorge Cwik, */ /* * Changes: * Pedro Roque : Fast Retransmit/Recovery. * Two receive queues. * Retransmit queue handled by TCP. * Better retransmit timer handling. * New congestion avoidance. * Header prediction. * Variable renaming. * * Eric : Fast Retransmit. * Randy Scott : MSS option defines. * Eric Schenk : Fixes to slow start algorithm. * Eric Schenk : Yet another double ACK bug. * Eric Schenk : Delayed ACK bug fixes. * Eric Schenk : Floyd style fast retrans war avoidance. * David S. Miller : Don't allow zero congestion window. * Eric Schenk : Fix retransmitter so that it sends * next packet on ack of previous packet. * Andi Kleen : Moved open_request checking here * and process RSTs for open_requests. * Andi Kleen : Better prune_queue, and other fixes. * Andrey Savochkin: Fix RTT measurements in the presence of * timestamps. * Andrey Savochkin: Check sequence numbers correctly when * removing SACKs due to in sequence incoming * data segments. * Andi Kleen: Make sure we never ack data there is not * enough room for. Also make this condition * a fatal error if it might still happen. * Andi Kleen: Add tcp_measure_rcv_mss to make * connections with MSS #include #include #include #include #include #include #include int sysctl_tcp_timestamps __read_mostly = 1; int sysctl_tcp_window_scaling __read_mostly = 1; int sysctl_tcp_sack __read_mostly = 1; int sysctl_tcp_fack __read_mostly = 1; int sysctl_tcp_reordering __read_mostly = TCP_FASTRETRANS_THRESH; int sysctl_tcp_ecn __read_mostly; int sysctl_tcp_dsack __read_mostly = 1; int sysctl_tcp_app_win __read_mostly = 31; int sysctl_tcp_adv_win_scale __read_mostly = 2; int sysctl_tcp_stdurg __read_mostly; int sysctl_tcp_rfc1337 __read_mostly; int sysctl_tcp_max_orphans __read_mostly = NR_FILE; int sysctl_tcp_frto __read_mostly = 2; int sysctl_tcp_frto_response __read_mostly; int sysctl_tcp_nometrics_save __read_mostly; int sysctl_tcp_moderate_rcvbuf __read_mostly = 1; int sysctl_tcp_abc __read_mostly; #define FLAG_DATA 0x01 /* Incoming frame contained data. */ #define FLAG_WIN_UPDATE 0x02 /* Incoming ACK was a window update. */ #define FLAG_DATA_ACKED 0x04 /* This ACK acknowledged new data. */ #define FLAG_RETRANS_DATA_ACKED 0x08 /* "" "" some of which was retransmitted. */ #define FLAG_SYN_ACKED 0x10 /* This ACK acknowledged SYN. */ #define FLAG_DATA_SACKED 0x20 /* New SACK. */ #define FLAG_ECE 0x40 /* ECE in this ACK */ #define FLAG_DATA_LOST 0x80 /* SACK detected data lossage. */ #define FLAG_SLOWPATH 0x100 /* Do not skip RFC checks for window update.*/ #define FLAG_ONLY_ORIG_SACKED 0x200 /* SACKs only non-rexmit sent before RTO */ #define FLAG_SND_UNA_ADVANCED 0x400 /* Snd_una was changed (!= FLAG_DATA_ACKED) */ #define FLAG_DSACKING_ACK 0x800 /* SACK blocks contained DSACK info */ #define FLAG_NONHEAD_RETRANS_ACKED 0x1000 /* Non-head rexmitted data was ACKed */ #define FLAG_ACKED (FLAG_DATA_ACKED|FLAG_SYN_ACKED) #define FLAG_NOT_DUP (FLAG_DATA|FLAG_WIN_UPDATE|FLAG_ACKED) #define FLAG_CA_ALERT (FLAG_DATA_SACKED|FLAG_ECE) #define FLAG_FORWARD_PROGRESS (FLAG_ACKED|FLAG_DATA_SACKED) #define FLAG_ANY_PROGRESS (FLAG_FORWARD_PROGRESS|FLAG_SND_UNA_ADVANCED) #define IsSackFrto() (sysctl_tcp_frto == 0x2) #define TCP_REMNANT (TCP_FLAG_FIN|TCP_FLAG_URG|TCP_FLAG_SYN|TCP_FLAG_PSH) #define TCP_HP_BITS (~(TCP_RESERVED_BITS|TCP_FLAG_PSH)) /* Adapt the MSS value used to make delayed ack decision to the * real world. */ static void tcp_measure_rcv_mss(struct sock *sk, const struct sk_buff *skb) { struct inet_connection_sock *icsk = inet_csk(sk); const unsigned int lss = icsk->icsk_ack.last_seg_size; unsigned int len; icsk->icsk_ack.last_seg_size = 0; /* skb->len may jitter because of SACKs, even if peer * sends good full-sized frames. */ len = skb_shinfo(skb)->gso_size ?: skb->len; if (len >= icsk->icsk_ack.rcv_mss) { icsk->icsk_ack.rcv_mss = len; } else { /* Otherwise, we make more careful check taking into account, * that SACKs block is variable. * * "len" is invariant segment length, including TCP header. */ len += skb->data - skb_transport_header(skb); if (len >= TCP_MIN_RCVMSS + sizeof(struct tcphdr) || /* If PSH is not set, packet should be * full sized, provided peer TCP is not badly broken. * This observation (if it is correct 8)) allows * to handle super-low mtu links fairly. */ (len >= TCP_MIN_MSS + sizeof(struct tcphdr) && !(tcp_flag_word(tcp_hdr(skb)) & TCP_REMNANT))) { /* Subtract also invariant (if peer is RFC compliant), * tcp header plus fixed timestamp option length. * Resulting "len" is MSS free of SACK jitter. */ len -= tcp_sk(sk)->tcp_header_len; icsk->icsk_ack.last_seg_size = len; if (len == lss) { icsk->icsk_ack.rcv_mss = len; return; } } if (icsk->icsk_ack.pending & ICSK_ACK_PUSHED) icsk->icsk_ack.pending |= ICSK_ACK_PUSHED2; icsk->icsk_ack.pending |= ICSK_ACK_PUSHED; } } static void tcp_incr_quickack(struct sock *sk) { struct inet_connection_sock *icsk = inet_csk(sk); unsigned quickacks = tcp_sk(sk)->rcv_wnd / (2 * icsk->icsk_ack.rcv_mss); if (quickacks==0) quickacks=2; if (quickacks > icsk->icsk_ack.quick) icsk->icsk_ack.quick = min(quickacks, TCP_MAX_QUICKACKS); } void tcp_enter_quickack_mode(struct sock *sk) { struct inet_connection_sock *icsk = inet_csk(sk); tcp_incr_quickack(sk); icsk->icsk_ack.pingpong = 0; icsk->icsk_ack.ato = TCP_ATO_MIN; } /* Send ACKs quickly, if "quick" count is not exhausted * and the session is not interactive. */ static inline int tcp_in_quickack_mode(const struct sock *sk) { const struct inet_connection_sock *icsk = inet_csk(sk); return icsk->icsk_ack.quick && !icsk->icsk_ack.pingpong; } static inline void TCP_ECN_queue_cwr(struct tcp_sock *tp) { if (tp->ecn_flags&TCP_ECN_OK) tp->ecn_flags |= TCP_ECN_QUEUE_CWR; } static inline void TCP_ECN_accept_cwr(struct tcp_sock *tp, struct sk_buff *skb) { if (tcp_hdr(skb)->cwr) tp->ecn_flags &= ~TCP_ECN_DEMAND_CWR; } static inline void TCP_ECN_withdraw_cwr(struct tcp_sock *tp) { tp->ecn_flags &= ~TCP_ECN_DEMAND_CWR; } static inline void TCP_ECN_check_ce(struct tcp_sock *tp, struct sk_buff *skb) { if (tp->ecn_flags&TCP_ECN_OK) { if (INET_ECN_is_ce(TCP_SKB_CB(skb)->flags)) tp->ecn_flags |= TCP_ECN_DEMAND_CWR; /* Funny extension: if ECT is not set on a segment, * it is surely retransmit. It is not in ECN RFC, * but Linux follows this rule. */ else if (INET_ECN_is_not_ect((TCP_SKB_CB(skb)->flags))) tcp_enter_quickack_mode((struct sock *)tp); } } static inline void TCP_ECN_rcv_synack(struct tcp_sock *tp, struct tcphdr *th) { if ((tp->ecn_flags&TCP_ECN_OK) && (!th->ece || th->cwr)) tp->ecn_flags &= ~TCP_ECN_OK; } static inline void TCP_ECN_rcv_syn(struct tcp_sock *tp, struct tcphdr *th) { if ((tp->ecn_flags&TCP_ECN_OK) && (!th->ece || !th->cwr)) tp->ecn_flags &= ~TCP_ECN_OK; } static inline int TCP_ECN_rcv_ecn_echo(struct tcp_sock *tp, struct tcphdr *th) { if (th->ece && !th->syn && (tp->ecn_flags&TCP_ECN_OK)) return 1; return 0; } /* Buffer size and advertised window tuning. * * 1. Tuning sk->sk_sndbuf, when connection enters established state. */ static void tcp_fixup_sndbuf(struct sock *sk) { int sndmem = tcp_sk(sk)->rx_opt.mss_clamp + MAX_TCP_HEADER + 16 + sizeof(struct sk_buff); if (sk->sk_sndbuf < 3 * sndmem) sk->sk_sndbuf = min(3 * sndmem, sysctl_tcp_wmem[2]); } /* 2. Tuning advertised window (window_clamp, rcv_ssthresh) * * All tcp_full_space() is split to two parts: "network" buffer, allocated * forward and advertised in receiver window (tp->rcv_wnd) and * "application buffer", required to isolate scheduling/application * latencies from network. * window_clamp is maximal advertised window. It can be less than * tcp_full_space(), in this case tcp_full_space() - window_clamp * is reserved for "application" buffer. The less window_clamp is * the smoother our behaviour from viewpoint of network, but the lower * throughput and the higher sensitivity of the connection to losses. 8) * * rcv_ssthresh is more strict window_clamp used at "slow start" * phase to predict further behaviour of this connection. * It is used for two goals: * - to enforce header prediction at sender, even when application * requires some significant "application buffer". It is check #1. * - to prevent pruning of receive queue because of misprediction * of receiver window. Check #2. * * The scheme does not work when sender sends good segments opening * window and then starts to feed us spaghetti. But it should work * in common situations. Otherwise, we have to rely on queue collapsing. */ /* Slow part of check#2. */ static int __tcp_grow_window(const struct sock *sk, const struct sk_buff *skb) { struct tcp_sock *tp = tcp_sk(sk); /* Optimize this! */ int truesize = tcp_win_from_space(skb->truesize)/2; int window = tcp_win_from_space(sysctl_tcp_rmem[2])/2; while (tp->rcv_ssthresh <= window) { if (truesize <= skb->len) return 2 * inet_csk(sk)->icsk_ack.rcv_mss; truesize >>= 1; window >>= 1; } return 0; } static void tcp_grow_window(struct sock *sk, struct sk_buff *skb) { struct tcp_sock *tp = tcp_sk(sk); /* Check #1 */ if (tp->rcv_ssthresh < tp->window_clamp && (int)tp->rcv_ssthresh < tcp_space(sk) && !tcp_memory_pressure) { int incr; /* Check #2. Increase window, if skb with such overhead * will fit to rcvbuf in future. */ if (tcp_win_from_space(skb->truesize) <= skb->len) incr = 2*tp->advmss; else incr = __tcp_grow_window(sk, skb); if (incr) { tp->rcv_ssthresh = min(tp->rcv_ssthresh + incr, tp->window_clamp); inet_csk(sk)->icsk_ack.quick |= 1; } } } /* 3. Tuning rcvbuf, when connection enters established state. */ static void tcp_fixup_rcvbuf(struct sock *sk) { struct tcp_sock *tp = tcp_sk(sk); int rcvmem = tp->advmss + MAX_TCP_HEADER + 16 + sizeof(struct sk_buff); /* Try to select rcvbuf so that 4 mss-sized segments * will fit to window and corresponding skbs will fit to our rcvbuf. * (was 3; 4 is minimum to allow fast retransmit to work.) */ while (tcp_win_from_space(rcvmem) < tp->advmss) rcvmem += 128; if (sk->sk_rcvbuf < 4 * rcvmem) sk->sk_rcvbuf = min(4 * rcvmem, sysctl_tcp_rmem[2]); } /* 4. Try to fixup all. It is made immediately after connection enters * established state. */ static void tcp_init_buffer_space(struct sock *sk) { struct tcp_sock *tp = tcp_sk(sk); int maxwin; if (!(sk->sk_userlocks & SOCK_RCVBUF_LOCK)) tcp_fixup_rcvbuf(sk); if (!(sk->sk_userlocks & SOCK_SNDBUF_LOCK)) tcp_fixup_sndbuf(sk); tp->rcvq_space.space = tp->rcv_wnd; maxwin = tcp_full_space(sk); if (tp->window_clamp >= maxwin) { tp->window_clamp = maxwin; if (sysctl_tcp_app_win && maxwin > 4 * tp->advmss) tp->window_clamp = max(maxwin - (maxwin >> sysctl_tcp_app_win), 4 * tp->advmss); } /* Force reservation of one segment. */ if (sysctl_tcp_app_win && tp->window_clamp > 2 * tp->advmss && tp->window_clamp + tp->advmss > maxwin) tp->window_clamp = max(2 * tp->advmss, maxwin - tp->advmss); tp->rcv_ssthresh = min(tp->rcv_ssthresh, tp->window_clamp); tp->snd_cwnd_stamp = tcp_time_stamp; } /* 5. Recalculate window clamp after socket hit its memory bounds. */ static void tcp_clamp_window(struct sock *sk) { struct tcp_sock *tp = tcp_sk(sk); struct inet_connection_sock *icsk = inet_csk(sk); icsk->icsk_ack.quick = 0; if (sk->sk_rcvbuf < sysctl_tcp_rmem[2] && !(sk->sk_userlocks & SOCK_RCVBUF_LOCK) && !tcp_memory_pressure && atomic_read(&tcp_memory_allocated) < sysctl_tcp_mem[0]) { sk->sk_rcvbuf = min(atomic_read(&sk->sk_rmem_alloc), sysctl_tcp_rmem[2]); } if (atomic_read(&sk->sk_rmem_alloc) > sk->sk_rcvbuf) tp->rcv_ssthresh = min(tp->window_clamp, 2U*tp->advmss); } /* Initialize RCV_MSS value. * RCV_MSS is an our guess about MSS used by the peer. * We haven't any direct information about the MSS. * It's better to underestimate the RCV_MSS rather than overestimate. * Overestimations make us ACKing less frequently than needed. * Underestimations are more easy to detect and fix by tcp_measure_rcv_mss(). */ void tcp_initialize_rcv_mss(struct sock *sk) { struct tcp_sock *tp = tcp_sk(sk); unsigned int hint = min_t(unsigned int, tp->advmss, tp->mss_cache); hint = min(hint, tp->rcv_wnd/2); hint = min(hint, TCP_MIN_RCVMSS); hint = max(hint, TCP_MIN_MSS); inet_csk(sk)->icsk_ack.rcv_mss = hint; } /* Receiver "autotuning" code. * * The algorithm for RTT estimation w/o timestamps is based on * Dynamic Right-Sizing (DRS) by Wu Feng and Mike Fisk of LANL. * * * More detail on this code can be found at * , * though this reference is out of date. A new paper * is pending. */ static void tcp_rcv_rtt_update(struct tcp_sock *tp, u32 sample, int win_dep) { u32 new_sample = tp->rcv_rtt_est.rtt; long m = sample; if (m == 0) m = 1; if (new_sample != 0) { /* If we sample in larger samples in the non-timestamp * case, we could grossly overestimate the RTT especially * with chatty applications or bulk transfer apps which * are stalled on filesystem I/O. * * Also, since we are only going for a minimum in the * non-timestamp case, we do not smooth things out * else with timestamps disabled convergence takes too * long. */ if (!win_dep) { m -= (new_sample >> 3); new_sample += m; } else if (m < new_sample) new_sample = m << 3; } else { /* No previous measure. */ new_sample = m << 3; } if (tp->rcv_rtt_est.rtt != new_sample) tp->rcv_rtt_est.rtt = new_sample; } static inline void tcp_rcv_rtt_measure(struct tcp_sock *tp) { if (tp->rcv_rtt_est.time == 0) goto new_measure; if (before(tp->rcv_nxt, tp->rcv_rtt_est.seq)) return; tcp_rcv_rtt_update(tp, jiffies - tp->rcv_rtt_est.time, 1); new_measure: tp->rcv_rtt_est.seq = tp->rcv_nxt + tp->rcv_wnd; tp->rcv_rtt_est.time = tcp_time_stamp; } static inline void tcp_rcv_rtt_measure_ts(struct sock *sk, const struct sk_buff *skb) { struct tcp_sock *tp = tcp_sk(sk); if (tp->rx_opt.rcv_tsecr && (TCP_SKB_CB(skb)->end_seq - TCP_SKB_CB(skb)->seq >= inet_csk(sk)->icsk_ack.rcv_mss)) tcp_rcv_rtt_update(tp, tcp_time_stamp - tp->rx_opt.rcv_tsecr, 0); } /* * This function should be called every time data is copied to user space. * It calculates the appropriate TCP receive buffer space. */ void tcp_rcv_space_adjust(struct sock *sk) { struct tcp_sock *tp = tcp_sk(sk); int time; int space; if (tp->rcvq_space.time == 0) goto new_measure; time = tcp_time_stamp - tp->rcvq_space.time; if (time < (tp->rcv_rtt_est.rtt >> 3) || tp->rcv_rtt_est.rtt == 0) return; space = 2 * (tp->copied_seq - tp->rcvq_space.seq); space = max(tp->rcvq_space.space, space); if (tp->rcvq_space.space != space) { int rcvmem; tp->rcvq_space.space = space; if (sysctl_tcp_moderate_rcvbuf && !(sk->sk_userlocks & SOCK_RCVBUF_LOCK)) { int new_clamp = space; /* Receive space grows, normalize in order to * take into account packet headers and sk_buff * structure overhead. */ space /= tp->advmss; if (!space) space = 1; rcvmem = (tp->advmss + MAX_TCP_HEADER + 16 + sizeof(struct sk_buff)); while (tcp_win_from_space(rcvmem) < tp->advmss) rcvmem += 128; space *= rcvmem; space = min(space, sysctl_tcp_rmem[2]); if (space > sk->sk_rcvbuf) { sk->sk_rcvbuf = space; /* Make the window clamp follow along. */ tp->window_clamp = new_clamp; } } } new_measure: tp->rcvq_space.seq = tp->copied_seq; tp->rcvq_space.time = tcp_time_stamp; } /* There is something which you must keep in mind when you analyze the * behavior of the tp->ato delayed ack timeout interval. When a * connection starts up, we want to ack as quickly as possible. The * problem is that "good" TCP's do slow start at the beginning of data * transmission. The means that until we send the first few ACK's the * sender will sit on his end and only queue most of his data, because * he can only send snd_cwnd unacked packets at any given time. For * each ACK we send, he increments snd_cwnd and transmits more of his * queue. -DaveM */ static void tcp_event_data_recv(struct sock *sk, struct sk_buff *skb) { struct tcp_sock *tp = tcp_sk(sk); struct inet_connection_sock *icsk = inet_csk(sk); u32 now; inet_csk_schedule_ack(sk); tcp_measure_rcv_mss(sk, skb); tcp_rcv_rtt_measure(tp); now = tcp_time_stamp; if (!icsk->icsk_ack.ato) { /* The _first_ data packet received, initialize * delayed ACK engine. */ tcp_incr_quickack(sk); icsk->icsk_ack.ato = TCP_ATO_MIN; } else { int m = now - icsk->icsk_ack.lrcvtime; if (m <= TCP_ATO_MIN/2) { /* The fastest case is the first. */ icsk->icsk_ack.ato = (icsk->icsk_ack.ato >> 1) + TCP_ATO_MIN / 2; } else if (m < icsk->icsk_ack.ato) { icsk->icsk_ack.ato = (icsk->icsk_ack.ato >> 1) + m; if (icsk->icsk_ack.ato > icsk->icsk_rto) icsk->icsk_ack.ato = icsk->icsk_rto; } else if (m > icsk->icsk_rto) { /* Too long gap. Apparently sender failed to * restart window, so that we send ACKs quickly. */ tcp_incr_quickack(sk); sk_stream_mem_reclaim(sk); } } icsk->icsk_ack.lrcvtime = now; TCP_ECN_check_ce(tp, skb); if (skb->len >= 128) tcp_grow_window(sk, skb); } static u32 tcp_rto_min(struct sock *sk) { struct dst_entry *dst = __sk_dst_get(sk); u32 rto_min = TCP_RTO_MIN; if (dst && dst_metric_locked(dst, RTAX_RTO_MIN)) rto_min = dst->metrics[RTAX_RTO_MIN-1]; return rto_min; } /* Called to compute a smoothed rtt estimate. The data fed to this * routine either comes from timestamps, or from segments that were * known _not_ to have been retransmitted [see Karn/Partridge * Proceedings SIGCOMM 87]. The algorithm is from the SIGCOMM 88 * piece by Van Jacobson. * NOTE: the next three routines used to be one big routine. * To save cycles in the RFC 1323 implementation it was better to break * it up into three procedures. -- erics */ static void tcp_rtt_estimator(struct sock *sk, const __u32 mrtt) { struct tcp_sock *tp = tcp_sk(sk); long m = mrtt; /* RTT */ /* The following amusing code comes from Jacobson's * article in SIGCOMM '88. Note that rtt and mdev * are scaled versions of rtt and mean deviation. * This is designed to be as fast as possible * m stands for "measurement". * * On a 1990 paper the rto value is changed to: * RTO = rtt + 4 * mdev * * Funny. This algorithm seems to be very broken. * These formulae increase RTO, when it should be decreased, increase * too slowly, when it should be increased quickly, decrease too quickly * etc. I guess in BSD RTO takes ONE value, so that it is absolutely * does not matter how to _calculate_ it. Seems, it was trap * that VJ failed to avoid. 8) */ if (m == 0) m = 1; if (tp->srtt != 0) { m -= (tp->srtt >> 3); /* m is now error in rtt est */ tp->srtt += m; /* rtt = 7/8 rtt + 1/8 new */ if (m < 0) { m = -m; /* m is now abs(error) */ m -= (tp->mdev >> 2); /* similar update on mdev */ /* This is similar to one of Eifel findings. * Eifel blocks mdev updates when rtt decreases. * This solution is a bit different: we use finer gain * for mdev in this case (alpha*beta). * Like Eifel it also prevents growth of rto, * but also it limits too fast rto decreases, * happening in pure Eifel. */ if (m > 0) m >>= 3; } else { m -= (tp->mdev >> 2); /* similar update on mdev */ } tp->mdev += m; /* mdev = 3/4 mdev + 1/4 new */ if (tp->mdev > tp->mdev_max) { tp->mdev_max = tp->mdev; if (tp->mdev_max > tp->rttvar) tp->rttvar = tp->mdev_max; } if (after(tp->snd_una, tp->rtt_seq)) { if (tp->mdev_max < tp->rttvar) tp->rttvar -= (tp->rttvar-tp->mdev_max)>>2; tp->rtt_seq = tp->snd_nxt; tp->mdev_max = tcp_rto_min(sk); } } else { /* no previous measure. */ tp->srtt = m<<3; /* take the measured time to be rtt */ tp->mdev = m<<1; /* make sure rto = 3*rtt */ tp->mdev_max = tp->rttvar = max(tp->mdev, tcp_rto_min(sk)); tp->rtt_seq = tp->snd_nxt; } } /* Calculate rto without backoff. This is the second half of Van Jacobson's * routine referred to above. */ static inline void tcp_set_rto(struct sock *sk) { const struct tcp_sock *tp = tcp_sk(sk); /* Old crap is replaced with new one. 8) * * More seriously: * 1. If rtt variance happened to be less 50msec, it is hallucination. * It cannot be less due to utterly erratic ACK generation made * at least by solaris and freebsd. "Erratic ACKs" has _nothing_ * to do with delayed acks, because at cwnd>2 true delack timeout * is invisible. Actually, Linux-2.4 also generates erratic * ACKs in some circumstances. */ inet_csk(sk)->icsk_rto = (tp->srtt >> 3) + tp->rttvar; /* 2. Fixups made earlier cannot be right. * If we do not estimate RTO correctly without them, * all the algo is pure shit and should be replaced * with correct one. It is exactly, which we pretend to do. */ } /* NOTE: clamping at TCP_RTO_MIN is not required, current algo * guarantees that rto is higher. */ static inline void tcp_bound_rto(struct sock *sk) { if (inet_csk(sk)->icsk_rto > TCP_RTO_MAX) inet_csk(sk)->icsk_rto = TCP_RTO_MAX; } /* Save metrics learned by this TCP session. This function is called only, when TCP finishes successfully i.e. when it enters TIME-WAIT or goes from LAST-ACK to CLOSE. */ void tcp_update_metrics(struct sock *sk) { struct tcp_sock *tp = tcp_sk(sk); struct dst_entry *dst = __sk_dst_get(sk); if (sysctl_tcp_nometrics_save) return; dst_confirm(dst); if (dst && (dst->flags&DST_HOST)) { const struct inet_connection_sock *icsk = inet_csk(sk); int m; if (icsk->icsk_backoff || !tp->srtt) { /* This session failed to estimate rtt. Why? * Probably, no packets returned in time. * Reset our results. */ if (!(dst_metric_locked(dst, RTAX_RTT))) dst->metrics[RTAX_RTT-1] = 0; return; } m = dst_metric(dst, RTAX_RTT) - tp->srtt; /* If newly calculated rtt larger than stored one, * store new one. Otherwise, use EWMA. Remember, * rtt overestimation is always better than underestimation. */ if (!(dst_metric_locked(dst, RTAX_RTT))) { if (m <= 0) dst->metrics[RTAX_RTT-1] = tp->srtt; else dst->metrics[RTAX_RTT-1] -= (m>>3); } if (!(dst_metric_locked(dst, RTAX_RTTVAR))) { if (m < 0) m = -m; /* Scale deviation to rttvar fixed point */ m >>= 1; if (m < tp->mdev) m = tp->mdev; if (m >= dst_metric(dst, RTAX_RTTVAR)) dst->metrics[RTAX_RTTVAR-1] = m; else dst->metrics[RTAX_RTTVAR-1] -= (dst->metrics[RTAX_RTTVAR-1] - m)>>2; } if (tp->snd_ssthresh >= 0xFFFF) { /* Slow start still did not finish. */ if (dst_metric(dst, RTAX_SSTHRESH) && !dst_metric_locked(dst, RTAX_SSTHRESH) && (tp->snd_cwnd >> 1) > dst_metric(dst, RTAX_SSTHRESH)) dst->metrics[RTAX_SSTHRESH-1] = tp->snd_cwnd >> 1; if (!dst_metric_locked(dst, RTAX_CWND) && tp->snd_cwnd > dst_metric(dst, RTAX_CWND)) dst->metrics[RTAX_CWND-1] = tp->snd_cwnd; } else if (tp->snd_cwnd > tp->snd_ssthresh && icsk->icsk_ca_state == TCP_CA_Open) { /* Cong. avoidance phase, cwnd is reliable. */ if (!dst_metric_locked(dst, RTAX_SSTHRESH)) dst->metrics[RTAX_SSTHRESH-1] = max(tp->snd_cwnd >> 1, tp->snd_ssthresh); if (!dst_metric_locked(dst, RTAX_CWND)) dst->metrics[RTAX_CWND-1] = (dst->metrics[RTAX_CWND-1] + tp->snd_cwnd) >> 1; } else { /* Else slow start did not finish, cwnd is non-sense, ssthresh may be also invalid. */ if (!dst_metric_locked(dst, RTAX_CWND)) dst->metrics[RTAX_CWND-1] = (dst->metrics[RTAX_CWND-1] + tp->snd_ssthresh) >> 1; if (dst->metrics[RTAX_SSTHRESH-1] && !dst_metric_locked(dst, RTAX_SSTHRESH) && tp->snd_ssthresh > dst->metrics[RTAX_SSTHRESH-1]) dst->metrics[RTAX_SSTHRESH-1] = tp->snd_ssthresh; } if (!dst_metric_locked(dst, RTAX_REORDERING)) { if (dst->metrics[RTAX_REORDERING-1] < tp->reordering && tp->reordering != sysctl_tcp_reordering) dst->metrics[RTAX_REORDERING-1] = tp->reordering; } } } /* Numbers are taken from RFC3390. * * John Heffner states: * * The RFC specifies a window of no more than 4380 bytes * unless 2*MSS > 4380. Reading the pseudocode in the RFC * is a bit misleading because they use a clamp at 4380 bytes * rather than use a multiplier in the relevant range. */ __u32 tcp_init_cwnd(struct tcp_sock *tp, struct dst_entry *dst) { __u32 cwnd = (dst ? dst_metric(dst, RTAX_INITCWND) : 0); if (!cwnd) { if (tp->mss_cache > 1460) cwnd = 2; else cwnd = (tp->mss_cache > 1095) ? 3 : 4; } return min_t(__u32, cwnd, tp->snd_cwnd_clamp); } /* Set slow start threshold and cwnd not falling to slow start */ void tcp_enter_cwr(struct sock *sk, const int set_ssthresh) { struct tcp_sock *tp = tcp_sk(sk); const struct inet_connection_sock *icsk = inet_csk(sk); tp->prior_ssthresh = 0; tp->bytes_acked = 0; if (icsk->icsk_ca_state < TCP_CA_CWR) { tp->undo_marker = 0; if (set_ssthresh) tp->snd_ssthresh = icsk->icsk_ca_ops->ssthresh(sk); tp->snd_cwnd = min(tp->snd_cwnd, tcp_packets_in_flight(tp) + 1U); tp->snd_cwnd_cnt = 0; tp->high_seq = tp->snd_nxt; tp->snd_cwnd_stamp = tcp_time_stamp; TCP_ECN_queue_cwr(tp); tcp_set_ca_state(sk, TCP_CA_CWR); } } /* * 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) { struct tcp_sock *tp = tcp_sk(sk); struct dst_entry *dst = __sk_dst_get(sk); if (dst == NULL) goto reset; dst_confirm(dst); if (dst_metric_locked(dst, RTAX_CWND)) tp->snd_cwnd_clamp = dst_metric(dst, RTAX_CWND); if (dst_metric(dst, RTAX_SSTHRESH)) { tp->snd_ssthresh = dst_metric(dst, RTAX_SSTHRESH); if (tp->snd_ssthresh > tp->snd_cwnd_clamp) tp->snd_ssthresh = tp->snd_cwnd_clamp; } if (dst_metric(dst, RTAX_REORDERING) && tp->reordering != dst_metric(dst, RTAX_REORDERING)) { tcp_disable_fack(tp); tp->reordering = dst_metric(dst, RTAX_REORDERING); } if (dst_metric(dst, RTAX_RTT) == 0) goto reset; if (!tp->srtt && dst_metric(dst, RTAX_RTT) < (TCP_TIMEOUT_INIT << 3)) goto reset; /* Initial rtt is determined from SYN,SYN-ACK. * The segment is small and rtt may appear much * less than real one. Use per-dst memory * to make it more realistic. * * A bit of theory. RTT is time passed after "normal" sized packet * is sent until it is ACKed. In normal circumstances sending small * packets force peer to delay ACKs and calculation is correct too. * The algorithm is adaptive and, provided we follow specs, it * NEVER underestimate RTT. BUT! If peer tries to make some clever * tricks sort of "quick acks" for time long enough to decrease RTT * to low value, and then abruptly stops to do it and starts to delay * ACKs, wait for troubles. */ if (dst_metric(dst, RTAX_RTT) > tp->srtt) { tp->srtt = dst_metric(dst, RTAX_RTT); tp->rtt_seq = tp->snd_nxt; } if (dst_metric(dst, RTAX_RTTVAR) > tp->mdev) { tp->mdev = dst_metric(dst, RTAX_RTTVAR); tp->mdev_max = tp->rttvar = max(tp->mdev, TCP_RTO_MIN); } tcp_set_rto(sk); tcp_bound_rto(sk); if (inet_csk(sk)->icsk_rto < TCP_TIMEOUT_INIT && !tp->rx_opt.saw_tstamp) goto reset; tp->snd_cwnd = tcp_init_cwnd(tp, dst); tp->snd_cwnd_stamp = tcp_time_stamp; return; reset: /* Play conservative. If timestamps are not * supported, TCP will fail to recalculate correct * rtt, if initial rto is too small. FORGET ALL AND RESET! */ if (!tp->rx_opt.saw_tstamp && tp->srtt) { tp->srtt = 0; tp->mdev = tp->mdev_max = tp->rttvar = TCP_TIMEOUT_INIT; inet_csk(sk)->icsk_rto = TCP_TIMEOUT_INIT; } } static void tcp_update_reordering(struct sock *sk, const int metric, const int ts) { struct tcp_sock *tp = tcp_sk(sk); if (metric > tp->reordering) { tp->reordering = min(TCP_MAX_REORDERING, metric); /* This exciting event is worth to be remembered. 8) */ if (ts) NET_INC_STATS_BH(LINUX_MIB_TCPTSREORDER); else if (tcp_is_reno(tp)) NET_INC_STATS_BH(LINUX_MIB_TCPRENOREORDER); else if (tcp_is_fack(tp)) NET_INC_STATS_BH(LINUX_MIB_TCPFACKREORDER); else NET_INC_STATS_BH(LINUX_MIB_TCPSACKREORDER); #if FASTRETRANS_DEBUG > 1 printk(KERN_DEBUG "Disorder%d %d %u f%u s%u rr%d\n", tp->rx_opt.sack_ok, inet_csk(sk)->icsk_ca_state, tp->reordering, tp->fackets_out, tp->sacked_out, tp->undo_marker ? tp->undo_retrans : 0); #endif tcp_disable_fack(tp); } } /* This procedure tags the retransmission queue when SACKs arrive. * * We have three tag bits: SACKED(S), RETRANS(R) and LOST(L). * Packets in queue with these bits set are counted in variables * sacked_out, retrans_out and lost_out, correspondingly. * * Valid combinations are: * Tag InFlight Description * 0 1 - orig segment is in flight. * S 0 - nothing flies, orig reached receiver. * L 0 - nothing flies, orig lost by net. * R 2 - both orig and retransmit are in flight. * L|R 1 - orig is lost, retransmit is in flight. * S|R 1 - orig reached receiver, retrans is still in flight. * (L|S|R is logically valid, it could occur when L|R is sacked, * but it is equivalent to plain S and code short-curcuits it to S. * L|S is logically invalid, it would mean -1 packet in flight 8)) * * These 6 states form finite state machine, controlled by the following events: * 1. New ACK (+SACK) arrives. (tcp_sacktag_write_queue()) * 2. Retransmission. (tcp_retransmit_skb(), tcp_xmit_retransmit_queue()) * 3. Loss detection event of one of three flavors: * A. Scoreboard estimator decided the packet is lost. * A'. Reno "three dupacks" marks head of queue lost. * A''. Its FACK modfication, head until snd.fack is lost. * B. SACK arrives sacking data transmitted after never retransmitted * hole was sent out. * C. SACK arrives sacking SND.NXT at the moment, when the * segment was retransmitted. * 4. D-SACK added new rule: D-SACK changes any tag to S. * * It is pleasant to note, that state diagram turns out to be commutative, * so that we are allowed not to be bothered by order of our actions, * when multiple events arrive simultaneously. (see the function below). * * Reordering detection. * -------------------- * Reordering metric is maximal distance, which a packet can be displaced * in packet stream. With SACKs we can estimate it: * * 1. SACK fills old hole and the corresponding segment was not * ever retransmitted -> reordering. Alas, we cannot use it * when segment was retransmitted. * 2. The last flaw is solved with D-SACK. D-SACK arrives * 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 reporting * SACK reneging when it should advance SND.UNA. Such SACK block this is * perfectly valid, however, in light of RFC2018 which explicitly states * that "SACK block MUST reflect the newest segment. Even if the newest * segment is going to be discarded ...", not that it looks very clever * in case of head skb. Due to potentional receiver driven attacks, we * choose to avoid immediate execution of a walk in write queue due to * reneging and defer head skb's loss recovery to standard loss recovery * procedure that will eventually trigger (nothing forbids us doing this). * * 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); } /* Check for lost retransmit. This superb idea is borrowed from "ratehalving". * Event "C". Later note: FACK people cheated me again 8), we have to account * for reordering! Ugly, but should help. * * Search retransmitted skbs from write_queue that were sent when snd_nxt was * less than what is now known to be received by the other end (derived from * SACK blocks by the caller). Also calculate the lowest snd_nxt among the * remaining retransmitted skbs to avoid some costly processing per ACKs. */ static int tcp_mark_lost_retrans(struct sock *sk, u32 received_upto) { struct tcp_sock *tp = tcp_sk(sk); struct sk_buff *skb; int flag = 0; int cnt = 0; u32 new_low_seq = tp->snd_nxt; tcp_for_write_queue(skb, sk) { u32 ack_seq = TCP_SKB_CB(skb)->ack_seq; if (skb == tcp_send_head(sk)) break; if (cnt == tp->retrans_out) break; if (!after(TCP_SKB_CB(skb)->end_seq, tp->snd_una)) continue; if (!(TCP_SKB_CB(skb)->sacked & TCPCB_SACKED_RETRANS)) continue; if (after(received_upto, ack_seq) && (tcp_is_fack(tp) || !before(received_upto, ack_seq + tp->reordering * tp->mss_cache))) { TCP_SKB_CB(skb)->sacked &= ~TCPCB_SACKED_RETRANS; tp->retrans_out -= tcp_skb_pcount(skb); /* clear lost hint */ tp->retransmit_skb_hint = NULL; if (!(TCP_SKB_CB(skb)->sacked & (TCPCB_LOST|TCPCB_SACKED_ACKED))) { tp->lost_out += tcp_skb_pcount(skb); TCP_SKB_CB(skb)->sacked |= TCPCB_LOST; flag |= FLAG_DATA_SACKED; NET_INC_STATS_BH(LINUX_MIB_TCPLOSTRETRANSMIT); } } else { if (before(ack_seq, new_low_seq)) new_low_seq = ack_seq; cnt += tcp_skb_pcount(skb); } } if (tp->retrans_out) tp->lost_retrans_low = new_low_seq; return flag; } 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) { u32 start_seq_0 = ntohl(get_unaligned(&sp[0].start_seq)); u32 end_seq_0 = ntohl(get_unaligned(&sp[0].end_seq)); int dup_sack = 0; if (before(start_seq_0, TCP_SKB_CB(ack_skb)->ack_seq)) { dup_sack = 1; 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)); u32 start_seq_1 = ntohl(get_unaligned(&sp[1].start_seq)); if (!after(end_seq_0, end_seq_1) && !before(start_seq_0, start_seq_1)) { dup_sack = 1; tcp_dsack_seen(tp); NET_INC_STATS_BH(LINUX_MIB_TCPDSACKOFORECV); } } /* D-SACK for already forgotten data... Do dumb counting. */ if (dup_sack && !after(end_seq_0, prior_snd_una) && after(end_seq_0, tp->undo_marker)) tp->undo_retrans--; return dup_sack; } /* Check if skb is fully within the SACK block. In presence of GSO skbs, * the incoming SACK may not exactly match but we can find smaller MSS * aligned portion of it that matches. Therefore we might need to fragment * which may fail and creates some hassle (caller must handle error case * returns). */ int tcp_match_skb_to_sack(struct sock *sk, struct sk_buff *skb, u32 start_seq, u32 end_seq) { int in_sack, err; unsigned int pkt_len; in_sack = !after(start_seq, TCP_SKB_CB(skb)->seq) && !before(end_seq, TCP_SKB_CB(skb)->end_seq); if (tcp_skb_pcount(skb) > 1 && !in_sack && after(TCP_SKB_CB(skb)->end_seq, start_seq)) { in_sack = !after(start_seq, TCP_SKB_CB(skb)->seq); if (!in_sack) pkt_len = start_seq - TCP_SKB_CB(skb)->seq; else pkt_len = end_seq - TCP_SKB_CB(skb)->seq; err = tcp_fragment(sk, skb, pkt_len, skb_shinfo(skb)->gso_size); if (err < 0) return err; } return in_sack; } static int tcp_sacktag_write_queue(struct sock *sk, struct sk_buff *ack_skb, u32 prior_snd_una) { const struct inet_connection_sock *icsk = inet_csk(sk); struct tcp_sock *tp = tcp_sk(sk); unsigned char *ptr = (skb_transport_header(ack_skb) + TCP_SKB_CB(ack_skb)->sacked); struct tcp_sack_block_wire *sp = (struct tcp_sack_block_wire *)(ptr+2); struct sk_buff *cached_skb; int num_sacks = (ptr[1] - TCPOLEN_SACK_BASE)>>3; int reord = tp->packets_out; int prior_fackets; u32 highest_sack_end_seq = tp->lost_retrans_low; int flag = 0; int found_dup_sack = 0; int cached_fack_count; int i; int first_sack_index; if (!tp->sacked_out) { if (WARN_ON(tp->fackets_out)) tp->fackets_out = 0; tp->highest_sack = tp->snd_una; } prior_fackets = tp->fackets_out; found_dup_sack = tcp_check_dsack(tp, ack_skb, sp, num_sacks, prior_snd_una); if (found_dup_sack) flag |= FLAG_DSACKING_ACK; /* Eliminate too old ACKs, but take into * account more or less fresh ones, they can * contain valid SACK info. */ if (before(TCP_SKB_CB(ack_skb)->ack_seq, prior_snd_una - tp->max_window)) return 0; /* SACK fastpath: * if the only SACK change is the increase of the end_seq of * the first block then only apply that SACK block * and use retrans queue hinting otherwise slowpath */ flag = 1; for (i = 0; i < num_sacks; i++) { __be32 start_seq = sp[i].start_seq; __be32 end_seq = sp[i].end_seq; if (i == 0) { if (tp->recv_sack_cache[i].start_seq != start_seq) flag = 0; } else { if ((tp->recv_sack_cache[i].start_seq != start_seq) || (tp->recv_sack_cache[i].end_seq != end_seq)) flag = 0; } tp->recv_sack_cache[i].start_seq = start_seq; tp->recv_sack_cache[i].end_seq = end_seq; } /* Clear the rest of the cache sack blocks so they won't match mistakenly. */ for (; i < ARRAY_SIZE(tp->recv_sack_cache); i++) { tp->recv_sack_cache[i].start_seq = 0; tp->recv_sack_cache[i].end_seq = 0; } first_sack_index = 0; if (flag) num_sacks = 1; else { int j; tp->fastpath_skb_hint = NULL; /* order SACK blocks to allow in order walk of the retrans queue */ for (i = num_sacks-1; i > 0; i--) { for (j = 0; j < i; j++){ if (after(ntohl(sp[j].start_seq), ntohl(sp[j+1].start_seq))){ struct tcp_sack_block_wire tmp; tmp = sp[j]; sp[j] = sp[j+1]; sp[j+1] = tmp; /* Track where the first SACK block goes to */ if (j == first_sack_index) first_sack_index = j+1; } } } } /* clear flag as used for different purpose in following code */ flag = 0; /* Use SACK fastpath hint if valid */ cached_skb = tp->fastpath_skb_hint; cached_fack_count = tp->fastpath_cnt_hint; if (!cached_skb) { cached_skb = tcp_write_queue_head(sk); cached_fack_count = 0; } for (i=0; istart_seq); __u32 end_seq = ntohl(sp->end_seq); 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 { /* Don't count olds caused by ACK reordering */ if ((TCP_SKB_CB(ack_skb)->ack_seq != tp->snd_una) && !after(end_seq, tp->snd_una)) continue; NET_INC_STATS_BH(LINUX_MIB_TCPSACKDISCARD); } continue; } skb = cached_skb; fack_count = cached_fack_count; /* Event "B" in the comment above. */ if (after(end_seq, tp->high_seq)) flag |= FLAG_DATA_LOST; tcp_for_write_queue_from(skb, sk) { int in_sack; u8 sacked; if (skb == tcp_send_head(sk)) break; cached_skb = skb; cached_fack_count = fack_count; if (i == first_sack_index) { tp->fastpath_skb_hint = skb; tp->fastpath_cnt_hint = fack_count; } /* The retransmission queue is always in order, so * we can short-circuit the walk early. */ if (!before(TCP_SKB_CB(skb)->seq, end_seq)) break; in_sack = tcp_match_skb_to_sack(sk, skb, start_seq, end_seq); if (in_sack < 0) break; fack_count += tcp_skb_pcount(skb); sacked = TCP_SKB_CB(skb)->sacked; /* Account D-SACK for retransmitted packet. */ if ((dup_sack && in_sack) && (sacked & TCPCB_RETRANS) && after(TCP_SKB_CB(skb)->end_seq, tp->undo_marker)) tp->undo_retrans--; /* The frame is ACKed. */ if (!after(TCP_SKB_CB(skb)->end_seq, tp->snd_una)) { if (sacked&TCPCB_RETRANS) { if ((dup_sack && in_sack) && (sacked&TCPCB_SACKED_ACKED)) reord = min(fack_count, reord); } else { /* If it was in a hole, we detected reordering. */ if (fack_count < prior_fackets && !(sacked&TCPCB_SACKED_ACKED)) reord = min(fack_count, reord); } /* Nothing to do; acked frame is about to be dropped. */ continue; } if (!in_sack) continue; if (!(sacked&TCPCB_SACKED_ACKED)) { if (sacked & TCPCB_SACKED_RETRANS) { /* If the segment is not tagged as lost, * we do not clear RETRANS, believing * that retransmission is still in flight. */ if (sacked & TCPCB_LOST) { TCP_SKB_CB(skb)->sacked &= ~(TCPCB_LOST|TCPCB_SACKED_RETRANS); tp->lost_out -= tcp_skb_pcount(skb); tp->retrans_out -= tcp_skb_pcount(skb); /* clear lost hint */ tp->retransmit_skb_hint = NULL; } } else { /* New sack for not retransmitted frame, * which was in hole. It is reordering. */ if (!(sacked & TCPCB_RETRANS) && fack_count < prior_fackets) reord = min(fack_count, reord); if (sacked & TCPCB_LOST) { TCP_SKB_CB(skb)->sacked &= ~TCPCB_LOST; tp->lost_out -= tcp_skb_pcount(skb); /* clear lost hint */ tp->retransmit_skb_hint = NULL; } /* SACK enhanced F-RTO detection. * Set flag if and only if non-rexmitted * segments below frto_highmark are * SACKed (RFC4138; Appendix B). * Clearing correct due to in-order walk */ if (after(end_seq, tp->frto_highmark)) { flag &= ~FLAG_ONLY_ORIG_SACKED; } else { if (!(sacked & TCPCB_RETRANS)) flag |= FLAG_ONLY_ORIG_SACKED; } } TCP_SKB_CB(skb)->sacked |= TCPCB_SACKED_ACKED; flag |= FLAG_DATA_SACKED; tp->sacked_out += tcp_skb_pcount(skb); if (fack_count > tp->fackets_out) tp->fackets_out = fack_count; if (after(TCP_SKB_CB(skb)->seq, tp->highest_sack)) { tp->highest_sack = TCP_SKB_CB(skb)->seq; highest_sack_end_seq = TCP_SKB_CB(skb)->end_seq; } } else { if (dup_sack && (sacked&TCPCB_RETRANS)) reord = min(fack_count, reord); } /* D-SACK. We can detect redundant retransmission * in S|R and plain R frames and clear it. * undo_retrans is decreased above, L|R frames * are accounted above as well. */ if (dup_sack && (TCP_SKB_CB(skb)->sacked&TCPCB_SACKED_RETRANS)) { TCP_SKB_CB(skb)->sacked &= ~TCPCB_SACKED_RETRANS; tp->retrans_out -= tcp_skb_pcount(skb); tp->retransmit_skb_hint = NULL; } } } if (tp->retrans_out && after(highest_sack_end_seq, tp->lost_retrans_low) && icsk->icsk_ca_state == TCP_CA_Recovery) flag |= tcp_mark_lost_retrans(sk, highest_sack_end_seq); tcp_verify_left_out(tp); if ((reord < tp->fackets_out) && icsk->icsk_ca_state != TCP_CA_Loss && (!tp->frto_highmark || after(tp->snd_una, tp->frto_highmark))) tcp_update_reordering(sk, ((tp->fackets_out + 1) - reord), 0); #if FASTRETRANS_DEBUG > 0 BUG_TRAP((int)tp->sacked_out >= 0); BUG_TRAP((int)tp->lost_out >= 0); BUG_TRAP((int)tp->retrans_out >= 0); BUG_TRAP((int)tcp_packets_in_flight(tp) >= 0); #endif return flag; } /* If we receive more dupacks than we expected counting segments * in assumption of absent reordering, interpret this as reordering. * The only another reason could be bug in receiver TCP. */ static void tcp_check_reno_reordering(struct sock *sk, const int addend) { struct tcp_sock *tp = tcp_sk(sk); u32 holes; holes = max(tp->lost_out, 1U); holes = min(holes, tp->packets_out); if ((tp->sacked_out + holes) > tp->packets_out) { tp->sacked_out = tp->packets_out - holes; tcp_update_reordering(sk, tp->packets_out + addend, 0); } } /* Emulate SACKs for SACKless connection: account for a new dupack. */ static void tcp_add_reno_sack(struct sock *sk) { struct tcp_sock *tp = tcp_sk(sk); tp->sacked_out++; tcp_check_reno_reordering(sk, 0); tcp_verify_left_out(tp); } /* Account for ACK, ACKing some data in Reno Recovery phase. */ static void tcp_remove_reno_sacks(struct sock *sk, int acked) { struct tcp_sock *tp = tcp_sk(sk); if (acked > 0) { /* One ACK acked hole. The rest eat duplicate ACKs. */ if (acked-1 >= tp->sacked_out) tp->sacked_out = 0; else tp->sacked_out -= acked-1; } tcp_check_reno_reordering(sk, acked); tcp_verify_left_out(tp); } static inline void tcp_reset_reno_sack(struct tcp_sock *tp) { tp->sacked_out = 0; } /* F-RTO can only be used if TCP has never retransmitted anything other than * head (SACK enhanced variant from Appendix B of RFC4138 is more robust here) */ int tcp_use_frto(struct sock *sk) { const struct tcp_sock *tp = tcp_sk(sk); struct sk_buff *skb; if (!sysctl_tcp_frto) return 0; if (IsSackFrto()) return 1; /* Avoid expensive walking of rexmit queue if possible */ if (tp->retrans_out > 1) return 0; skb = tcp_write_queue_head(sk); skb = tcp_write_queue_next(sk, skb); /* Skips head */ tcp_for_write_queue_from(skb, sk) { if (skb == tcp_send_head(sk)) break; if (TCP_SKB_CB(skb)->sacked&TCPCB_RETRANS) return 0; /* Short-circuit when first non-SACKed skb has been checked */ if (!(TCP_SKB_CB(skb)->sacked&TCPCB_SACKED_ACKED)) break; } return 1; } /* RTO occurred, but do not yet enter Loss state. Instead, defer RTO * recovery a bit and use heuristics in tcp_process_frto() to detect if * the RTO was spurious. Only clear SACKED_RETRANS of the head here to * keep retrans_out counting accurate (with SACK F-RTO, other than head * may still have that bit set); TCPCB_LOST and remaining SACKED_RETRANS * bits are handled if the Loss state is really to be entered (in * tcp_enter_frto_loss). * * Do like tcp_enter_loss() would; when RTO expires the second time it * does: * "Reduce ssthresh if it has not yet been made inside this window." */ void tcp_enter_frto(struct sock *sk) { const struct inet_connection_sock *icsk = inet_csk(sk); struct tcp_sock *tp = tcp_sk(sk); struct sk_buff *skb; if ((!tp->frto_counter && icsk->icsk_ca_state <= TCP_CA_Disorder) || tp->snd_una == tp->high_seq || ((icsk->icsk_ca_state == TCP_CA_Loss || tp->frto_counter) && !icsk->icsk_retransmits)) { tp->prior_ssthresh = tcp_current_ssthresh(sk); /* Our state is too optimistic in ssthresh() call because cwnd * is not reduced until tcp_enter_frto_loss() when previous FRTO * recovery has not yet completed. Pattern would be this: RTO, * Cumulative ACK, RTO (2xRTO for the same segment does not end * up here twice). * RFC4138 should be more specific on what to do, even though * RTO is quite unlikely to occur after the first Cumulative ACK * due to back-off and complexity of triggering events ... */ if (tp->frto_counter) { u32 stored_cwnd; stored_cwnd = tp->snd_cwnd; tp->snd_cwnd = 2; tp->snd_ssthresh = icsk->icsk_ca_ops->ssthresh(sk); tp->snd_cwnd = stored_cwnd; } else { tp->snd_ssthresh = icsk->icsk_ca_ops->ssthresh(sk); } /* ... in theory, cong.control module could do "any tricks" in * ssthresh(), which means that ca_state, lost bits and lost_out * counter would have to be faked before the call occurs. We * consider that too expensive, unlikely and hacky, so modules * using these in ssthresh() must deal these incompatibility * issues if they receives CA_EVENT_FRTO and frto_counter != 0 */ tcp_ca_event(sk, CA_EVENT_FRTO); } tp->undo_marker = tp->snd_una; tp->undo_retrans = 0; skb = tcp_write_queue_head(sk); if (TCP_SKB_CB(skb)->sacked & TCPCB_RETRANS) tp->undo_marker = 0; if (TCP_SKB_CB(skb)->sacked & TCPCB_SACKED_RETRANS) { TCP_SKB_CB(skb)->sacked &= ~TCPCB_SACKED_RETRANS; tp->retrans_out -= tcp_skb_pcount(skb); } tcp_verify_left_out(tp); /* Earlier loss recovery underway (see RFC4138; Appendix B). * The last condition is necessary at least in tp->frto_counter case. */ if (IsSackFrto() && (tp->frto_counter || ((1 << icsk->icsk_ca_state) & (TCPF_CA_Recovery|TCPF_CA_Loss))) && after(tp->high_seq, tp->snd_una)) { tp->frto_highmark = tp->high_seq; } else { tp->frto_highmark = tp->snd_nxt; } tcp_set_ca_state(sk, TCP_CA_Disorder); tp->high_seq = tp->snd_nxt; tp->frto_counter = 1; } /* Enter Loss state after F-RTO was applied. Dupack arrived after RTO, * which indicates that we should follow the traditional RTO recovery, * i.e. mark everything lost and do go-back-N retransmission. */ static void tcp_enter_frto_loss(struct sock *sk, int allowed_segments, int flag) { struct tcp_sock *tp = tcp_sk(sk); struct sk_buff *skb; tp->lost_out = 0; tp->retrans_out = 0; if (tcp_is_reno(tp)) tcp_reset_reno_sack(tp); tcp_for_write_queue(skb, sk) { if (skb == tcp_send_head(sk)) break; /* * Count the retransmission made on RTO correctly (only when * waiting for the first ACK and did not get it)... */ if ((tp->frto_counter == 1) && !(flag&FLAG_DATA_ACKED)) { /* For some reason this R-bit might get cleared? */ if (TCP_SKB_CB(skb)->sacked & TCPCB_SACKED_RETRANS) tp->retrans_out += tcp_skb_pcount(skb); /* ...enter this if branch just for the first segment */ flag |= FLAG_DATA_ACKED; } else { if (TCP_SKB_CB(skb)->sacked & TCPCB_RETRANS) tp->undo_marker = 0; TCP_SKB_CB(skb)->sacked &= ~(TCPCB_LOST|TCPCB_SACKED_RETRANS); } /* Don't lost mark skbs that were fwd transmitted after RTO */ if (!(TCP_SKB_CB(skb)->sacked&TCPCB_SACKED_ACKED) && !after(TCP_SKB_CB(skb)->end_seq, tp->frto_highmark)) { TCP_SKB_CB(skb)->sacked |= TCPCB_LOST; tp->lost_out += tcp_skb_pcount(skb); } } tcp_verify_left_out(tp); tp->snd_cwnd = tcp_packets_in_flight(tp) + allowed_segments; tp->snd_cwnd_cnt = 0; tp->snd_cwnd_stamp = tcp_time_stamp; tp->frto_counter = 0; tp->bytes_acked = 0; tp->reordering = min_t(unsigned int, tp->reordering, sysctl_tcp_reordering); tcp_set_ca_state(sk, TCP_CA_Loss); tp->high_seq = tp->frto_highmark; TCP_ECN_queue_cwr(tp); tcp_clear_retrans_hints_partial(tp); } static void tcp_clear_retrans_partial(struct tcp_sock *tp) { tp->retrans_out = 0; tp->lost_out = 0; tp->undo_marker = 0; tp->undo_retrans = 0; } void tcp_clear_retrans(struct tcp_sock *tp) { tcp_clear_retrans_partial(tp); tp->fackets_out = 0; tp->sacked_out = 0; } /* Enter Loss state. If "how" is not zero, forget all SACK information * and reset tags completely, otherwise preserve SACKs. If receiver * dropped its ofo queue, we will know this due to reneging detection. */ void tcp_enter_loss(struct sock *sk, int how) { const struct inet_connection_sock *icsk = inet_csk(sk); struct tcp_sock *tp = tcp_sk(sk); struct sk_buff *skb; /* Reduce ssthresh if it has not yet been made inside this window. */ if (icsk->icsk_ca_state <= TCP_CA_Disorder || tp->snd_una == tp->high_seq || (icsk->icsk_ca_state == TCP_CA_Loss && !icsk->icsk_retransmits)) { tp->prior_ssthresh = tcp_current_ssthresh(sk); tp->snd_ssthresh = icsk->icsk_ca_ops->ssthresh(sk); tcp_ca_event(sk, CA_EVENT_LOSS); } tp->snd_cwnd = 1; tp->snd_cwnd_cnt = 0; tp->snd_cwnd_stamp = tcp_time_stamp; tp->bytes_acked = 0; tcp_clear_retrans_partial(tp); if (tcp_is_reno(tp)) tcp_reset_reno_sack(tp); if (!how) { /* Push undo marker, if it was plain RTO and nothing * was retransmitted. */ tp->undo_marker = tp->snd_una; tcp_clear_retrans_hints_partial(tp); } else { tp->sacked_out = 0; tp->fackets_out = 0; tcp_clear_all_retrans_hints(tp); } tcp_for_write_queue(skb, sk) { if (skb == tcp_send_head(sk)) break; if (TCP_SKB_CB(skb)->sacked&TCPCB_RETRANS) tp->undo_marker = 0; TCP_SKB_CB(skb)->sacked &= (~TCPCB_TAGBITS)|TCPCB_SACKED_ACKED; if (!(TCP_SKB_CB(skb)->sacked&TCPCB_SACKED_ACKED) || how) { TCP_SKB_CB(skb)->sacked &= ~TCPCB_SACKED_ACKED; TCP_SKB_CB(skb)->sacked |= TCPCB_LOST; tp->lost_out += tcp_skb_pcount(skb); } } tcp_verify_left_out(tp); tp->reordering = min_t(unsigned int, tp->reordering, sysctl_tcp_reordering); tcp_set_ca_state(sk, TCP_CA_Loss); tp->high_seq = tp->snd_nxt; TCP_ECN_queue_cwr(tp); /* Abort FRTO algorithm if one is in progress */ tp->frto_counter = 0; } static int tcp_check_sack_reneging(struct sock *sk) { struct sk_buff *skb; /* If ACK arrived pointing to a remembered SACK, * it means that our remembered SACKs do not reflect * real state of receiver i.e. * receiver _host_ is heavily congested (or buggy). * Do processing similar to RTO timeout. */ if ((skb = tcp_write_queue_head(sk)) != NULL && (TCP_SKB_CB(skb)->sacked & TCPCB_SACKED_ACKED)) { struct inet_connection_sock *icsk = inet_csk(sk); NET_INC_STATS_BH(LINUX_MIB_TCPSACKRENEGING); tcp_enter_loss(sk, 1); icsk->icsk_retransmits++; tcp_retransmit_skb(sk, tcp_write_queue_head(sk)); inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS, icsk->icsk_rto, TCP_RTO_MAX); return 1; } return 0; } static inline int tcp_fackets_out(struct tcp_sock *tp) { 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) { return (tcp_time_stamp - TCP_SKB_CB(skb)->when > inet_csk(sk)->icsk_rto); } static inline int tcp_head_timedout(struct sock *sk) { struct tcp_sock *tp = tcp_sk(sk); return tp->packets_out && tcp_skb_timedout(sk, tcp_write_queue_head(sk)); } /* Linux NewReno/SACK/FACK/ECN state machine. * -------------------------------------- * * "Open" Normal state, no dubious events, fast path. * "Disorder" In all the respects it is "Open", * but requires a bit more attention. It is entered when * we see some SACKs or dupacks. It is split of "Open" * mainly to move some processing from fast path to slow one. * "CWR" CWND was reduced due to some Congestion Notification event. * It can be ECN, ICMP source quench, local device congestion. * "Recovery" CWND was reduced, we are fast-retransmitting. * "Loss" CWND was reduced due to RTO timeout or SACK reneging. * * tcp_fastretrans_alert() is entered: * - each incoming ACK, if state is not "Open" * - when arrived ACK is unusual, namely: * * SACK * * Duplicate ACK. * * ECN ECE. * * Counting packets in flight is pretty simple. * * in_flight = packets_out - left_out + retrans_out * * packets_out is SND.NXT-SND.UNA counted in packets. * * retrans_out is number of retransmitted segments. * * left_out is number of segments left network, but not ACKed yet. * * left_out = sacked_out + lost_out * * sacked_out: Packets, which arrived to receiver out of order * and hence not ACKed. With SACKs this number is simply * amount of SACKed data. Even without SACKs * it is easy to give pretty reliable estimate of this number, * counting duplicate ACKs. * * lost_out: Packets lost by network. TCP has no explicit * "loss notification" feedback from network (for now). * It means that this number can be only _guessed_. * Actually, it is the heuristics to predict lossage that * distinguishes different algorithms. * * F.e. after RTO, when all the queue is considered as lost, * lost_out = packets_out and in_flight = retrans_out. * * Essentially, we have now two algorithms counting * lost packets. * * FACK: It is the simplest heuristics. As soon as we decided * that something is lost, we decide that _all_ not SACKed * packets until the most forward SACK are lost. I.e. * lost_out = fackets_out - sacked_out and left_out = fackets_out. * It is absolutely correct estimate, if network does not reorder * packets. And it loses any connection to reality when reordering * takes place. We use FACK by default until reordering * is suspected on the path to this destination. * * NewReno: when Recovery is entered, we assume that one segment * is lost (classic Reno). While we are in Recovery and * a partial ACK arrives, we assume that one more packet * is lost (NewReno). This heuristics are the same in NewReno * and SACK. * * Imagine, that's all! Forget about all this shamanism about CWND inflation * deflation etc. CWND is real congestion window, never inflated, changes * only according to classic VJ rules. * * Really tricky (and requiring careful tuning) part of algorithm * is hidden in functions tcp_time_to_recover() and tcp_xmit_retransmit_queue(). * The first determines the moment _when_ we should reduce CWND and, * hence, slow down forward transmission. In fact, it determines the moment * when we decide that hole is caused by loss, rather than by a reorder. * * tcp_xmit_retransmit_queue() decides, _what_ we should retransmit to fill * holes, caused by lost packets. * * And the most logically complicated part of algorithm is undo * heuristics. We detect false retransmits due to both too early * fast retransmit (reordering) and underestimated RTO, analyzing * timestamps and D-SACKs. When we detect that some segments were * retransmitted by mistake and CWND reduction was wrong, we undo * window reduction and abort recovery phase. This logic is hidden * inside several functions named tcp_try_undo_. */ /* This function decides, when we should leave Disordered state * and enter Recovery phase, reducing congestion window. * * Main question: may we further continue forward transmission * with the same cwnd? */ static int tcp_time_to_recover(struct sock *sk) { struct tcp_sock *tp = tcp_sk(sk); __u32 packets_out; /* Do not perform any recovery during FRTO algorithm */ if (tp->frto_counter) return 0; /* Trick#1: The loss is proven. */ if (tp->lost_out) return 1; /* Not-A-Trick#2 : Classic rule... */ if (tcp_fackets_out(tp) > tp->reordering) return 1; /* Trick#3 : when we use RFC2988 timer restart, fast * retransmit can be triggered by timeout of queue head. */ if (tcp_head_timedout(sk)) return 1; /* Trick#4: It is still not OK... But will it be useful to delay * recovery more? */ packets_out = tp->packets_out; if (packets_out <= tp->reordering && tp->sacked_out >= max_t(__u32, packets_out/2, sysctl_tcp_reordering) && !tcp_may_send_now(sk)) { /* We have nothing to send. This connection is limited * either by receiver window or by application. */ return 1; } return 0; } /* RFC: This is from the original, I doubt that this is necessary at all: * clear xmit_retrans hint if seq of this skb is beyond hint. How could we * retransmitted past LOST markings in the first place? I'm not fully sure * about undo and end of connection cases, which can cause R without L? */ static void tcp_verify_retransmit_hint(struct tcp_sock *tp, struct sk_buff *skb) { if ((tp->retransmit_skb_hint != NULL) && before(TCP_SKB_CB(skb)->seq, TCP_SKB_CB(tp->retransmit_skb_hint)->seq)) tp->retransmit_skb_hint = NULL; } /* Mark head of queue up as lost. */ static void tcp_mark_head_lost(struct sock *sk, int packets) { struct tcp_sock *tp = tcp_sk(sk); struct sk_buff *skb; int cnt; BUG_TRAP(packets <= tp->packets_out); if (tp->lost_skb_hint) { skb = tp->lost_skb_hint; cnt = tp->lost_cnt_hint; } else { skb = tcp_write_queue_head(sk); cnt = 0; } tcp_for_write_queue_from(skb, sk) { if (skb == tcp_send_head(sk)) break; /* TODO: do this better */ /* this is not the most efficient way to do this... */ tp->lost_skb_hint = skb; tp->lost_cnt_hint = cnt; cnt += tcp_skb_pcount(skb); if (cnt > packets || after(TCP_SKB_CB(skb)->end_seq, tp->high_seq)) break; if (!(TCP_SKB_CB(skb)->sacked & (TCPCB_SACKED_ACKED|TCPCB_LOST))) { TCP_SKB_CB(skb)->sacked |= TCPCB_LOST; tp->lost_out += tcp_skb_pcount(skb); tcp_verify_retransmit_hint(tp, skb); } } tcp_verify_left_out(tp); } /* Account newly detected lost packet(s) */ static void tcp_update_scoreboard(struct sock *sk) { struct tcp_sock *tp = tcp_sk(sk); if (tcp_is_fack(tp)) { int lost = tp->fackets_out - tp->reordering; if (lost <= 0) lost = 1;