diff options
Diffstat (limited to 'net/ipv4/tcp_output.c')
| -rw-r--r-- | net/ipv4/tcp_output.c | 279 |
1 files changed, 145 insertions, 134 deletions
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 975f4472af29..0faacf9c419d 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -62,14 +62,13 @@ int sysctl_tcp_base_mss __read_mostly = 512; /* By default, RFC2861 behavior. */ int sysctl_tcp_slow_start_after_idle __read_mostly = 1; -static void update_send_head(struct sock *sk, struct tcp_sock *tp, - struct sk_buff *skb) +static void update_send_head(struct sock *sk, struct sk_buff *skb) { - sk->sk_send_head = skb->next; - if (sk->sk_send_head == (struct sk_buff *)&sk->sk_write_queue) - sk->sk_send_head = NULL; + struct tcp_sock *tp = tcp_sk(sk); + + tcp_advance_send_head(sk, skb); tp->snd_nxt = TCP_SKB_CB(skb)->end_seq; - tcp_packets_out_inc(sk, tp, skb); + tcp_packets_out_inc(sk, skb); } /* SND.NXT, if window was not shrunk. @@ -78,8 +77,10 @@ static void update_send_head(struct sock *sk, struct tcp_sock *tp, * Anything in between SND.UNA...SND.UNA+SND.WND also can be already * invalid. OK, let's make this for now: */ -static inline __u32 tcp_acceptable_seq(struct sock *sk, struct tcp_sock *tp) +static inline __u32 tcp_acceptable_seq(struct sock *sk) { + struct tcp_sock *tp = tcp_sk(sk); + if (!before(tp->snd_una+tp->snd_wnd, tp->snd_nxt)) return tp->snd_nxt; else @@ -198,7 +199,7 @@ void tcp_select_initial_window(int __space, __u32 mss, (*rcv_wscale) = 0; if (wscale_ok) { /* Set window scaling on max possible window - * See RFC1323 for an explanation of the limit to 14 + * See RFC1323 for an explanation of the limit to 14 */ space = max_t(u32, sysctl_tcp_rmem[2], sysctl_rmem_max); space = min_t(u32, space, *window_clamp); @@ -238,7 +239,7 @@ static u16 tcp_select_window(struct sock *sk) u32 new_win = __tcp_select_window(sk); /* Never shrink the offered window */ - if(new_win < cur_win) { + if (new_win < cur_win) { /* Danger Will Robinson! * Don't update rcv_wup/rcv_wnd here or else * we will not be able to advertise a zero @@ -289,10 +290,12 @@ static void tcp_build_and_update_options(__be32 *ptr, struct tcp_sock *tp, (TCPOPT_SACK << 8) | (TCPOLEN_SACK_BASE + (tp->rx_opt.eff_sacks * TCPOLEN_SACK_PERBLOCK))); - for(this_sack = 0; this_sack < tp->rx_opt.eff_sacks; this_sack++) { + + for (this_sack = 0; this_sack < tp->rx_opt.eff_sacks; this_sack++) { *ptr++ = htonl(sp[this_sack].start_seq); *ptr++ = htonl(sp[this_sack].end_seq); } + if (tp->rx_opt.dsack) { tp->rx_opt.dsack = 0; tp->rx_opt.eff_sacks--; @@ -337,7 +340,7 @@ static void tcp_syn_build_options(__be32 *ptr, int mss, int ts, int sack, */ *ptr++ = htonl((TCPOPT_MSS << 24) | (TCPOLEN_MSS << 16) | mss); if (ts) { - if(sack) + if (sack) *ptr++ = htonl((TCPOPT_SACK_PERM << 24) | (TCPOLEN_SACK_PERM << 16) | (TCPOPT_TIMESTAMP << 8) | @@ -349,7 +352,7 @@ static void tcp_syn_build_options(__be32 *ptr, int mss, int ts, int sack, TCPOLEN_TIMESTAMP); *ptr++ = htonl(tstamp); /* TSVAL */ *ptr++ = htonl(ts_recent); /* TSECR */ - } else if(sack) + } else if (sack) *ptr++ = htonl((TCPOPT_NOP << 24) | (TCPOPT_NOP << 16) | (TCPOPT_SACK_PERM << 8) | @@ -406,7 +409,7 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it, /* If congestion control is doing timestamping, we must * take such a timestamp before we potentially clone/copy. */ - if (icsk->icsk_ca_ops->rtt_sample) + if (icsk->icsk_ca_ops->flags & TCP_CONG_RTT_STAMP) __net_timestamp(skb); if (likely(clone_it)) { @@ -430,7 +433,7 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it, sysctl_flags = 0; if (unlikely(tcb->flags & TCPCB_FLAG_SYN)) { tcp_header_size = sizeof(struct tcphdr) + TCPOLEN_MSS; - if(sysctl_tcp_timestamps) { + if (sysctl_tcp_timestamps) { tcp_header_size += TCPOLEN_TSTAMP_ALIGNED; sysctl_flags |= SYSCTL_FLAG_TSTAMPS; } @@ -451,7 +454,7 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it, (tp->rx_opt.eff_sacks * TCPOLEN_SACK_PERBLOCK)); } - + if (tcp_packets_in_flight(tp) == 0) tcp_ca_event(sk, CA_EVENT_TX_START); @@ -465,11 +468,12 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it, tcp_header_size += TCPOLEN_MD5SIG_ALIGNED; #endif - th = (struct tcphdr *) skb_push(skb, tcp_header_size); - skb->h.th = th; + skb_push(skb, tcp_header_size); + skb_reset_transport_header(skb); skb_set_owner_w(skb, sk); /* Build TCP header and checksum it. */ + th = tcp_hdr(skb); th->source = inet->sport; th->dest = inet->dport; th->seq = htonl(tcb->seq); @@ -481,7 +485,7 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it, /* RFC1323: The window in SYN & SYN/ACK segments * is never scaled. */ - th->window = htons(tp->rcv_wnd); + th->window = htons(min(tp->rcv_wnd, 65535U)); } else { th->window = htons(tcp_select_window(sk)); } @@ -515,7 +519,7 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it, md5 ? &md5_hash_location : #endif NULL); - TCP_ECN_send(sk, tp, skb, tcp_header_size); + TCP_ECN_send(sk, skb, tcp_header_size); } #ifdef CONFIG_TCP_MD5SIG @@ -524,7 +528,7 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it, tp->af_specific->calc_md5_hash(md5_hash_location, md5, sk, NULL, NULL, - skb->h.th, + tcp_hdr(skb), sk->sk_protocol, skb->len); } @@ -545,7 +549,7 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it, if (likely(err <= 0)) return err; - tcp_enter_cwr(sk); + tcp_enter_cwr(sk, 1); return net_xmit_eval(err); @@ -555,7 +559,7 @@ static int tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, int clone_it, } -/* This routine just queue's the buffer +/* This routine just queue's the buffer * * NOTE: probe0 timer is not checked, do not forget tcp_push_pending_frames, * otherwise socket can stall. @@ -567,12 +571,8 @@ static void tcp_queue_skb(struct sock *sk, struct sk_buff *skb) /* Advance write_seq and place onto the write_queue. */ tp->write_seq = TCP_SKB_CB(skb)->end_seq; skb_header_release(skb); - __skb_queue_tail(&sk->sk_write_queue, skb); + tcp_add_write_queue_tail(sk, skb); sk_charge_skb(sk, skb); - - /* Queue it, remembering where we must start sending. */ - if (sk->sk_send_head == NULL) - sk->sk_send_head = skb; } static void tcp_set_skb_tso_segs(struct sock *sk, struct sk_buff *skb, unsigned int mss_now) @@ -597,7 +597,7 @@ static void tcp_set_skb_tso_segs(struct sock *sk, struct sk_buff *skb, unsigned /* Function to create two new TCP segments. Shrinks the given segment * to the specified size and appends a new segment with the rest of the - * packet to the list. This won't be called frequently, I hope. + * packet to the list. This won't be called frequently, I hope. * Remember, these are still headerless SKBs at this point. */ int tcp_fragment(struct sock *sk, struct sk_buff *skb, u32 len, unsigned int mss_now) @@ -610,7 +610,7 @@ int tcp_fragment(struct sock *sk, struct sk_buff *skb, u32 len, unsigned int mss BUG_ON(len > skb->len); - clear_all_retrans_hints(tp); + clear_all_retrans_hints(tp); nsize = skb_headlen(skb) - len; if (nsize < 0) nsize = 0; @@ -705,7 +705,7 @@ int tcp_fragment(struct sock *sk, struct sk_buff *skb, u32 len, unsigned int mss /* Link BUFF into the send queue. */ skb_header_release(buff); - __skb_append(skb, buff, &sk->sk_write_queue); + tcp_insert_write_queue_after(skb, buff, sk); return 0; } @@ -736,7 +736,7 @@ static void __pskb_trim_head(struct sk_buff *skb, int len) } skb_shinfo(skb)->nr_frags = k; - skb->tail = skb->data; + skb_reset_tail_pointer(skb); skb->data_len -= len; skb->len = skb->data_len; } @@ -821,7 +821,7 @@ void tcp_mtup_init(struct sock *sk) icsk->icsk_mtup.enabled = sysctl_tcp_mtu_probing > 1; icsk->icsk_mtup.search_high = tp->rx_opt.mss_clamp + sizeof(struct tcphdr) + - icsk->icsk_af_ops->net_header_len; + icsk->icsk_af_ops->net_header_len; icsk->icsk_mtup.search_low = tcp_mss_to_mtu(sk, sysctl_tcp_base_mss); icsk->icsk_mtup.probe_size = 0; } @@ -930,8 +930,9 @@ unsigned int tcp_current_mss(struct sock *sk, int large_allowed) /* Congestion window validation. (RFC2861) */ -static void tcp_cwnd_validate(struct sock *sk, struct tcp_sock *tp) +static void tcp_cwnd_validate(struct sock *sk) { + struct tcp_sock *tp = tcp_sk(sk); __u32 packets_out = tp->packets_out; if (packets_out >= tp->snd_cwnd) { @@ -943,7 +944,8 @@ static void tcp_cwnd_validate(struct sock *sk, struct tcp_sock *tp) if (tp->packets_out > tp->snd_cwnd_used) tp->snd_cwnd_used = tp->packets_out; - if ((s32)(tcp_time_stamp - tp->snd_cwnd_stamp) >= inet_csk(sk)->icsk_rto) + if (sysctl_tcp_slow_start_after_idle && + (s32)(tcp_time_stamp - tp->snd_cwnd_stamp) >= inet_csk(sk)->icsk_rto) tcp_cwnd_application_limited(sk); } } @@ -965,7 +967,8 @@ static inline unsigned int tcp_cwnd_test(struct tcp_sock *tp, struct sk_buff *sk u32 in_flight, cwnd; /* Don't be strict about the congestion window for the final FIN. */ - if (TCP_SKB_CB(skb)->flags & TCPCB_FLAG_FIN) + if ((TCP_SKB_CB(skb)->flags & TCPCB_FLAG_FIN) && + tcp_skb_pcount(skb) == 1) return 1; in_flight = tcp_packets_in_flight(tp); @@ -1007,7 +1010,7 @@ static inline int tcp_minshall_check(const struct tcp_sock *tp) */ static inline int tcp_nagle_check(const struct tcp_sock *tp, - const struct sk_buff *skb, + const struct sk_buff *skb, unsigned mss_now, int nonagle) { return (skb->len < mss_now && @@ -1032,8 +1035,10 @@ static inline int tcp_nagle_test(struct tcp_sock *tp, struct sk_buff *skb, if (nonagle & TCP_NAGLE_PUSH) return 1; - /* Don't use the nagle rule for urgent data (or for the final FIN). */ - if (tp->urg_mode || + /* Don't use the nagle rule for urgent data (or for the final FIN). + * Nagle can be ignored during F-RTO too (see RFC4138). + */ + if (tp->urg_mode || (tp->frto_counter == 2) || (TCP_SKB_CB(skb)->flags & TCPCB_FLAG_FIN)) return 1; @@ -1054,7 +1059,7 @@ static inline int tcp_snd_wnd_test(struct tcp_sock *tp, struct sk_buff *skb, uns return !after(end_seq, tp->snd_una + tp->snd_wnd); } -/* This checks if the data bearing packet SKB (usually sk->sk_send_head) +/* This checks if the data bearing packet SKB (usually tcp_send_head(sk)) * should be put on the wire right now. If so, it returns the number of * packets allowed by the congestion window. */ @@ -1077,15 +1082,10 @@ static unsigned int tcp_snd_test(struct sock *sk, struct sk_buff *skb, return cwnd_quota; } -static inline int tcp_skb_is_last(const struct sock *sk, - const struct sk_buff *skb) -{ - return skb->next == (struct sk_buff *)&sk->sk_write_queue; -} - -int tcp_may_send_now(struct sock *sk, struct tcp_sock *tp) +int tcp_may_send_now(struct sock *sk) { - struct sk_buff *skb = sk->sk_send_head; + struct tcp_sock *tp = tcp_sk(sk); + struct sk_buff *skb = tcp_send_head(sk); return (skb && tcp_snd_test(sk, skb, tcp_current_mss(sk, 1), @@ -1141,7 +1141,7 @@ static int tso_fragment(struct sock *sk, struct sk_buff *skb, unsigned int len, /* Link BUFF into the send queue. */ skb_header_release(buff); - __skb_append(skb, buff, &sk->sk_write_queue); + tcp_insert_write_queue_after(skb, buff, sk); return 0; } @@ -1151,8 +1151,9 @@ static int tso_fragment(struct sock *sk, struct sk_buff *skb, unsigned int len, * * This algorithm is from John Heffner. */ -static int tcp_tso_should_defer(struct sock *sk, struct tcp_sock *tp, struct sk_buff *skb) +static int tcp_tso_should_defer(struct sock *sk, struct sk_buff *skb) { + struct tcp_sock *tp = tcp_sk(sk); const struct inet_connection_sock *icsk = inet_csk(sk); u32 send_win, cong_win, limit, in_flight; @@ -1247,10 +1248,10 @@ static int tcp_mtu_probe(struct sock *sk) /* Have enough data in the send queue to probe? */ len = 0; - if ((skb = sk->sk_send_head) == NULL) + if ((skb = tcp_send_head(sk)) == NULL) return -1; while ((len += skb->len) < probe_size && !tcp_skb_is_last(sk, skb)) - skb = skb->next; + skb = tcp_write_queue_next(sk, skb); if (len < probe_size) return -1; @@ -1277,9 +1278,9 @@ static int tcp_mtu_probe(struct sock *sk) return -1; sk_charge_skb(sk, nskb); - skb = sk->sk_send_head; - __skb_insert(nskb, skb->prev, skb, &sk->sk_write_queue); - sk->sk_send_head = nskb; + skb = tcp_send_head(sk); + tcp_insert_write_queue_before(nskb, skb, sk); + tcp_advance_send_head(sk, skb); TCP_SKB_CB(nskb)->seq = TCP_SKB_CB(skb)->seq; TCP_SKB_CB(nskb)->end_seq = TCP_SKB_CB(skb)->seq + probe_size; @@ -1290,24 +1291,24 @@ static int tcp_mtu_probe(struct sock *sk) len = 0; while (len < probe_size) { - next = skb->next; + next = tcp_write_queue_next(sk, skb); copy = min_t(int, skb->len, probe_size - len); if (nskb->ip_summed) skb_copy_bits(skb, 0, skb_put(nskb, copy), copy); else nskb->csum = skb_copy_and_csum_bits(skb, 0, - skb_put(nskb, copy), copy, nskb->csum); + skb_put(nskb, copy), copy, nskb->csum); if (skb->len <= copy) { /* We've eaten all the data from this skb. * Throw it away. */ TCP_SKB_CB(nskb)->flags |= TCP_SKB_CB(skb)->flags; - __skb_unlink(skb, &sk->sk_write_queue); + tcp_unlink_write_queue(skb, sk); sk_stream_free_skb(sk, skb); } else { TCP_SKB_CB(nskb)->flags |= TCP_SKB_CB(skb)->flags & - ~(TCPCB_FLAG_FIN|TCPCB_FLAG_PSH); + ~(TCPCB_FLAG_FIN|TCPCB_FLAG_PSH); if (!skb_shinfo(skb)->nr_frags) { skb_pull(skb, copy); if (skb->ip_summed != CHECKSUM_PARTIAL) @@ -1331,7 +1332,7 @@ static int tcp_mtu_probe(struct sock *sk) /* Decrement cwnd here because we are sending * effectively two packets. */ tp->snd_cwnd--; - update_send_head(sk, tp, nskb); + update_send_head(sk, nskb); icsk->icsk_mtup.probe_size = tcp_mss_to_mtu(sk, nskb->len); tp->mtu_probe.probe_seq_start = TCP_SKB_CB(nskb)->seq; @@ -1375,7 +1376,7 @@ static int tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle) sent_pkts = 1; } - while ((skb = sk->sk_send_head)) { + while ((skb = tcp_send_head(sk))) { unsigned int limit; tso_segs = tcp_init_tso_segs(sk, skb, mss_now); @@ -1394,7 +1395,7 @@ static int tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle) nonagle : TCP_NAGLE_PUSH)))) break; } else { - if (tcp_tso_should_defer(sk, tp, skb)) + if (tcp_tso_should_defer(sk, skb)) break; } @@ -1423,31 +1424,31 @@ static int tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle) /* Advance the send_head. This one is sent out. * This call will increment packets_out. */ - update_send_head(sk, tp, skb); + update_send_head(sk, skb); tcp_minshall_update(tp, mss_now, skb); sent_pkts++; } if (likely(sent_pkts)) { - tcp_cwnd_validate(sk, tp); + tcp_cwnd_validate(sk); return 0; } - return !tp->packets_out && sk->sk_send_head; + return !tp->packets_out && tcp_send_head(sk); } /* Push out any pending frames which were held back due to * TCP_CORK or attempt at coalescing tiny packets. * The socket must be locked by the caller. */ -void __tcp_push_pending_frames(struct sock *sk, struct tcp_sock *tp, - unsigned int cur_mss, int nonagle) +void __tcp_push_pending_frames(struct sock *sk, unsigned int cur_mss, + int nonagle) { - struct sk_buff *skb = sk->sk_send_head; + struct sk_buff *skb = tcp_send_head(sk); if (skb) { if (tcp_write_xmit(sk, cur_mss, nonagle)) - tcp_check_probe_timer(sk, tp); + tcp_check_probe_timer(sk); } } @@ -1457,7 +1458,7 @@ void __tcp_push_pending_frames(struct sock *sk, struct tcp_sock *tp, void tcp_push_one(struct sock *sk, unsigned int mss_now) { struct tcp_sock *tp = tcp_sk(sk); - struct sk_buff *skb = sk->sk_send_head; + struct sk_buff *skb = tcp_send_head(sk); unsigned int tso_segs, cwnd_quota; BUG_ON(!skb || skb->len < mss_now); @@ -1491,8 +1492,8 @@ void tcp_push_one(struct sock *sk, unsigned int mss_now) TCP_SKB_CB(skb)->when = tcp_time_stamp; if (likely(!tcp_transmit_skb(sk, skb, 1, sk->sk_allocation))) { - update_send_head(sk, tp, skb); - tcp_cwnd_validate(sk, tp); + update_send_head(sk, skb); + tcp_cwnd_validate(sk); return; } } @@ -1500,7 +1501,7 @@ void tcp_push_one(struct sock *sk, unsigned int mss_now) /* This function returns the amount that we can raise the * usable window based on the following constraints - * + * * 1. The window can never be shrunk once it is offered (RFC 793) * 2. We limit memory per socket * @@ -1519,12 +1520,12 @@ void tcp_push_one(struct sock *sk, unsigned int mss_now) * side SWS prevention criteria. The problem is that under this rule * a stream of single byte packets will cause the right side of the * window to always advance by a single byte. - * + * * Of course, if the sender implements sender side SWS prevention * then this will not be a problem. - * + * * BSD seems to make the following compromise: - * + * * If the free space is less than the 1/4 of the maximum * space available and the free space is less than 1/2 mss, * then set the window to 0. @@ -1566,7 +1567,7 @@ u32 __tcp_select_window(struct sock *sk) int window; if (mss > full_space) - mss = full_space; + mss = full_space; if (free_space < full_space/2) { icsk->icsk_ack.quick = 0; @@ -1606,6 +1607,9 @@ u32 __tcp_select_window(struct sock *sk) */ if (window <= free_space - mss || window > free_space) window = (free_space/mss)*mss; + else if (mss == full_space && + free_space > window + full_space/2) + window = free_space; } return window; @@ -1615,7 +1619,7 @@ u32 __tcp_select_window(struct sock *sk) static void tcp_retrans_try_collapse(struct sock *sk, struct sk_buff *skb, int mss_now) { struct tcp_sock *tp = tcp_sk(sk); - struct sk_buff *next_skb = skb->next; + struct sk_buff *next_skb = tcp_write_queue_next(sk, skb); /* The first test we must make is that neither of these two * SKB's are still referenced by someone else. @@ -1625,7 +1629,7 @@ static void tcp_retrans_try_collapse(struct sock *sk, struct sk_buff *skb, int m u16 flags = TCP_SKB_CB(skb)->flags; /* Also punt if next skb has been SACK'd. */ - if(TCP_SKB_CB(next_skb)->sacked & TCPCB_SACKED_ACKED) + if (TCP_SKB_CB(next_skb)->sacked & TCPCB_SACKED_ACKED) return; /* Next skb is out of window. */ @@ -1647,9 +1651,11 @@ static void tcp_retrans_try_collapse(struct sock *sk, struct sk_buff *skb, int m clear_all_retrans_hints(tp); /* Ok. We will be able to collapse the packet. */ - __skb_unlink(next_skb, &sk->sk_write_queue); + tcp_unlink_write_queue(next_skb, sk); - memcpy(skb_put(skb, next_skb_size), next_skb->data, next_skb_size); + skb_copy_from_linear_data(next_skb, + skb_put(skb, next_skb_size), + next_skb_size); if (next_skb->ip_summed == CHECKSUM_PARTIAL) skb->ip_summed = CHECKSUM_PARTIAL; @@ -1690,9 +1696,9 @@ static void tcp_retrans_try_collapse(struct sock *sk, struct sk_buff *skb, int m } /* Do a simple retransmit without using the backoff mechanisms in - * tcp_timer. This is used for path mtu discovery. + * tcp_timer. This is used for path mtu discovery. * The socket is already locked here. - */ + */ void tcp_simple_retransmit(struct sock *sk) { const struct inet_connection_sock *icsk = inet_csk(sk); @@ -1701,8 +1707,10 @@ void tcp_simple_retransmit(struct sock *sk) unsigned int mss = tcp_current_mss(sk, 0); int lost = 0; - sk_stream_for_retrans_queue(skb, sk) { - if (skb->len > mss && + tcp_for_write_queue(skb, sk) { + if (skb == tcp_send_head(sk)) + break; + if (skb->len > mss && !(TCP_SKB_CB(skb)->sacked&TCPCB_SACKED_ACKED)) { if (TCP_SKB_CB(skb)->sacked&TCPCB_SACKED_RETRANS) { TCP_SKB_CB(skb)->sacked &= ~TCPCB_SACKED_RETRANS; @@ -1723,7 +1731,7 @@ void tcp_simple_retransmit(struct sock *sk) tcp_sync_left_out(tp); - /* Don't muck with the congestion window here. + /* Don't muck with the congestion window here. * Reason is that we do not increase amount of _data_ * in network, but units changed and effective * cwnd/ssthresh really reduced now. @@ -1746,7 +1754,7 @@ int tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb) { struct tcp_sock *tp = tcp_sk(sk); struct inet_connection_sock *icsk = inet_csk(sk); - unsigned int cur_mss = tcp_current_mss(sk, 0); + unsigned int cur_mss = tcp_current_mss(sk, 0); int err; /* Inconslusive MTU probe */ @@ -1783,13 +1791,13 @@ int tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb) } /* Collapse two adjacent packets if worthwhile and we can. */ - if(!(TCP_SKB_CB(skb)->flags & TCPCB_FLAG_SYN) && - (skb->len < (cur_mss >> 1)) && - (skb->next != sk->sk_send_head) && - (skb->next != (struct sk_buff *)&sk->sk_write_queue) && - (skb_shinfo(skb)->nr_frags == 0 && skb_shinfo(skb->next)->nr_frags == 0) && - (tcp_skb_pcount(skb) == 1 && tcp_skb_pcount(skb->next) == 1) && - (sysctl_tcp_retrans_collapse != 0)) + if (!(TCP_SKB_CB(skb)->flags & TCPCB_FLAG_SYN) && + (skb->len < (cur_mss >> 1)) && + (tcp_write_queue_next(sk, skb) != tcp_send_head(sk)) && + (!tcp_skb_is_last(sk, skb)) && + (skb_shinfo(skb)->nr_frags == 0 && skb_shinfo(tcp_write_queue_next(sk, skb))->nr_frags == 0) && + (tcp_skb_pcount(skb) == 1 && tcp_skb_pcount(tcp_write_queue_next(sk, skb)) == 1) && + (sysctl_tcp_retrans_collapse != 0)) tcp_retrans_try_collapse(sk, skb, cur_mss); if (inet_csk(sk)->icsk_af_ops->rebuild_header(sk)) @@ -1799,9 +1807,9 @@ int tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb) * retransmit when old data is attached. So strip it off * since it is cheap to do so and saves bytes on the network. */ - if(skb->len > 0 && - (TCP_SKB_CB(skb)->flags & TCPCB_FLAG_FIN) && - tp->snd_una == (TCP_SKB_CB(skb)->end_seq - 1)) { + if (skb->len > 0 && + (TCP_SKB_CB(skb)->flags & TCPCB_FLAG_FIN) && + tp->snd_una == (TCP_SKB_CB(skb)->end_seq - 1)) { if (!pskb_trim(skb, 0)) { TCP_SKB_CB(skb)->seq = TCP_SKB_CB(skb)->end_seq - 1; skb_shinfo(skb)->gso_segs = 1; @@ -1867,15 +1875,17 @@ void tcp_xmit_retransmit_queue(struct sock *sk) skb = tp->retransmit_skb_hint; packet_cnt = tp->retransmit_cnt_hint; }else{ - skb = sk->sk_write_queue.next; + skb = tcp_write_queue_head(sk); packet_cnt = 0; } /* First pass: retransmit lost packets. */ if (tp->lost_out) { - sk_stream_for_retrans_queue_from(skb, sk) { + 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; @@ -1901,8 +1911,7 @@ void tcp_xmit_retransmit_queue(struct sock *sk) else NET_INC_STATS_BH(LINUX_MIB_TCPSLOWSTARTRETRANS); - if (skb == - skb_peek(&sk->sk_write_queue)) + if (skb == tcp_write_queue_head(sk)) inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS, inet_csk(sk)->icsk_rto, TCP_RTO_MAX); @@ -1932,18 +1941,20 @@ void tcp_xmit_retransmit_queue(struct sock *sk) * segments to send. */ - if (tcp_may_send_now(sk, tp)) + if (tcp_may_send_now(sk)) return; if (tp->forward_skb_hint) { skb = tp->forward_skb_hint; packet_cnt = tp->forward_cnt_hint; } else{ - skb = sk->sk_write_queue.next; + skb = tcp_write_queue_head(sk); packet_cnt = 0; } - sk_stream_for_retrans_queue_from(skb, sk) { + tcp_for_write_queue_from(skb, sk) { + if (skb == tcp_send_head(sk)) + break; tp->forward_cnt_hint = packet_cnt; tp->forward_skb_hint = skb; @@ -1968,7 +1979,7 @@ void tcp_xmit_retransmit_queue(struct sock *sk) break; } - if (skb == skb_peek(&sk->sk_write_queue)) + if (skb == tcp_write_queue_head(sk)) inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS, inet_csk(sk)->icsk_rto, TCP_RTO_MAX); @@ -1983,17 +1994,17 @@ void tcp_xmit_retransmit_queue(struct sock *sk) */ void tcp_send_fin(struct sock *sk) { - struct tcp_sock *tp = tcp_sk(sk); - struct sk_buff *skb = skb_peek_tail(&sk->sk_write_queue); + struct tcp_sock *tp = tcp_sk(sk); + struct sk_buff *skb = tcp_write_queue_tail(sk); int mss_now; - + /* Optimization, tack on the FIN if we have a queue of * unsent frames. But be careful about outgoing SACKS * and IP options. */ mss_now = tcp_current_mss(sk, 1); - if (sk->sk_send_head != NULL) { + if (tcp_send_head(sk) != NULL) { TCP_SKB_CB(skb)->flags |= TCPCB_FLAG_FIN; TCP_SKB_CB(skb)->end_seq++; tp->write_seq++; @@ -2020,17 +2031,16 @@ void tcp_send_fin(struct sock *sk) TCP_SKB_CB(skb)->end_seq = TCP_SKB_CB(skb)->seq + 1; tcp_queue_skb(sk, skb); } - __tcp_push_pending_frames(sk, tp, mss_now, TCP_NAGLE_OFF); + __tcp_push_pending_frames(sk, mss_now, TCP_NAGLE_OFF); } /* We get here when a process closes a file descriptor (either due to * an explicit close() or as a byproduct of exit()'ing) and there * was unread data in the receive queue. This behavior is recommended - * by draft-ietf-tcpimpl-prob-03.txt section 3.10. -DaveM + * by RFC 2525, section 2.17. -DaveM */ void tcp_send_active_reset(struct sock *sk, gfp_t priority) { - struct tcp_sock *tp = tcp_sk(sk); struct sk_buff *skb; /* NOTE: No TCP options attached and we never retransmit this. */ @@ -2050,7 +2060,7 @@ void tcp_send_active_reset(struct sock *sk, gfp_t priority) skb_shinfo(skb)->gso_type = 0; /* Send it off. */ - TCP_SKB_CB(skb)->seq = tcp_acceptable_seq(sk, tp); + TCP_SKB_CB(skb)->seq = tcp_acceptable_seq(sk); TCP_SKB_CB(skb)->end_seq = TCP_SKB_CB(skb)->seq; TCP_SKB_CB(skb)->when = tcp_time_stamp; if (tcp_transmit_skb(sk, skb, 0, priority)) @@ -2066,7 +2076,7 @@ int tcp_send_synack(struct sock *sk) { struct sk_buff* skb; - skb = skb_peek(&sk->sk_write_queue); + skb = tcp_write_queue_head(sk); if (skb == NULL || !(TCP_SKB_CB(skb)->flags&TCPCB_FLAG_SYN)) { printk(KERN_DEBUG "tcp_send_synack: wrong queue state\n"); return -EFAULT; @@ -2076,9 +2086,9 @@ int tcp_send_synack(struct sock *sk) struct sk_buff *nskb = skb_copy(skb, GFP_ATOMIC); if (nskb == NULL) return -ENOMEM; - __skb_unlink(skb, &sk->sk_write_queue); + tcp_unlink_write_queue(skb, sk); skb_header_release(nskb); - __skb_queue_head(&sk->sk_write_queue, nskb); + __tcp_add_write_queue_head(sk, nskb); sk_stream_free_skb(sk, skb); sk_charge_skb(sk, nskb); skb = nskb; @@ -2128,8 +2138,10 @@ struct sk_buff * tcp_make_synack(struct sock *sk, struct dst_entry *dst, if (md5) tcp_header_size += TCPOLEN_MD5SIG_ALIGNED; #endif - skb->h.th = th = (struct tcphdr *) skb_push(skb, tcp_header_size); + skb_push(skb, tcp_header_size); + skb_reset_transport_header(skb); + th = tcp_hdr(skb); memset(th, 0, sizeof(struct tcphdr)); th->syn = 1; th->ack = 1; @@ -2145,21 +2157,21 @@ struct sk_buff * tcp_make_synack(struct sock *sk, struct dst_entry *dst, th->seq = htonl(TCP_SKB_CB(skb)->seq); th->ack_seq = htonl(tcp_rsk(req)->rcv_isn + 1); if (req->rcv_wnd == 0) { /* ignored for retransmitted syns */ - __u8 rcv_wscale; + __u8 rcv_wscale; /* Set this up on the first call only */ req->window_clamp = tp->window_clamp ? : dst_metric(dst, RTAX_WINDOW); /* tcp_full_space because it is guaranteed to be the first packet */ - tcp_select_initial_window(tcp_full_space(sk), + tcp_select_initial_window(tcp_full_space(sk), dst_metric(dst, RTAX_ADVMSS) - (ireq->tstamp_ok ? TCPOLEN_TSTAMP_ALIGNED : 0), &req->rcv_wnd, &req->window_clamp, ireq->wscale_ok, &rcv_wscale); - ireq->rcv_wscale = rcv_wscale; + ireq->rcv_wscale = rcv_wscale; } /* RFC1323: The window in SYN & SYN/ACK segments is never scaled. */ - th->window = htons(req->rcv_wnd); + th->window = htons(min(req->rcv_wnd, 65535U)); TCP_SKB_CB(skb)->when = tcp_time_stamp; tcp_syn_build_options((__be32 *)(th + 1), dst_metric(dst, RTAX_ADVMSS), ireq->tstamp_ok, @@ -2183,7 +2195,7 @@ struct sk_buff * tcp_make_synack(struct sock *sk, struct dst_entry *dst, tp->af_specific->calc_md5_hash(md5_hash_location, md5, NULL, dst, req, - skb->h.th, sk->sk_protocol, + tcp_hdr(skb), sk->sk_protocol, skb->len); } #endif @@ -2191,9 +2203,9 @@ struct sk_buff * tcp_make_synack(struct sock *sk, struct dst_entry *dst, return skb; } -/* +/* * Do all connect socket setups that can be done AF independent. - */ + */ static void tcp_connect_init(struct sock *sk) { struct dst_entry *dst = __sk_dst_get(sk); @@ -2250,7 +2262,7 @@ static void tcp_connect_init(struct sock *sk) /* * Build a SYN and send it off. - */ + */ int tcp_connect(struct sock *sk) { struct tcp_sock *tp = tcp_sk(sk); @@ -2266,7 +2278,7 @@ int tcp_connect(struct sock *sk) skb_reserve(buff, MAX_TCP_HEADER); TCP_SKB_CB(buff)->flags = TCPCB_FLAG_SYN; - TCP_ECN_send_syn(sk, tp, buff); + TCP_ECN_send_syn(sk, buff); TCP_SKB_CB(buff)->sacked = 0; skb_shinfo(buff)->gso_segs = 1; skb_shinfo(buff)->gso_size = 0; @@ -2280,7 +2292,7 @@ int tcp_connect(struct sock *sk) TCP_SKB_CB(buff)->when = tcp_time_stamp; tp->retrans_stamp = TCP_SKB_CB(buff)->when; skb_header_release(buff); - __skb_queue_tail(&sk->sk_write_queue, buff); + __tcp_add_write_queue_tail(sk, buff); sk_charge_skb(sk, buff); tp->packets_out += tcp_skb_pcount(buff); tcp_transmit_skb(sk, buff, 1, GFP_KERNEL); @@ -2358,7 +2370,6 @@ void tcp_send_ack(struct sock *sk) { /* If we have been reset, we may not send again. */ if (sk->sk_state != TCP_CLOSE) { - struct tcp_sock *tp = tcp_sk(sk); struct sk_buff *buff; /* We are not putting this on the write queue, so @@ -2384,7 +2395,7 @@ void tcp_send_ack(struct sock *sk) skb_shinfo(buff)->gso_type = 0; /* Send it off, this clears delayed acks for us. */ - TCP_SKB_CB(buff)->seq = TCP_SKB_CB(buff)->end_seq = tcp_acceptable_seq(sk, tp); + TCP_SKB_CB(buff)->seq = TCP_SKB_CB(buff)->end_seq = tcp_acceptable_seq(sk); TCP_SKB_CB(buff)->when = tcp_time_stamp; tcp_transmit_skb(sk, buff, 0, GFP_ATOMIC); } @@ -2408,7 +2419,7 @@ static int tcp_xmit_probe_skb(struct sock *sk, int urgent) /* We don't queue it, tcp_transmit_skb() sets ownership. */ skb = alloc_skb(MAX_TCP_HEADER, GFP_ATOMIC); - if (skb == NULL) + if (skb == NULL) return -1; /* Reserve space for headers and set control bits. */ @@ -2436,7 +2447,7 @@ int tcp_write_wakeup(struct sock *sk) struct tcp_sock *tp = tcp_sk(sk); struct sk_buff *skb; - if ((skb = sk->sk_send_head) != NULL && + if ((skb = tcp_send_head(sk)) != NULL && before(TCP_SKB_CB(skb)->seq, tp->snd_una+tp->snd_wnd)) { int err; unsigned int mss = tcp_current_mss(sk, 0); @@ -2462,7 +2473,7 @@ int tcp_write_wakeup(struct sock *sk) TCP_SKB_CB(skb)->when = tcp_time_stamp; err = tcp_transmit_skb(sk, skb, 1, GFP_ATOMIC); if (!err) { - update_send_head(sk, tp, skb); + update_send_head(sk, skb); } return err; } else { @@ -2486,7 +2497,7 @@ void tcp_send_probe0(struct sock *sk) err = tcp_write_wakeup(sk); - if (tp->packets_out || !sk->sk_send_head) { + if (tp->packets_out || !tcp_send_head(sk)) { /* Cancel probe timer, if it is not required. */ icsk->icsk_probes_out = 0; icsk->icsk_backoff = 0; @@ -2497,7 +2508,7 @@ void tcp_send_probe0(struct sock *sk) if (icsk->icsk_backoff < sysctl_tcp_retries2) icsk->icsk_backoff++; icsk->icsk_probes_out++; - inet_csk_reset_xmit_timer(sk, ICSK_TIME_PROBE0, + inet_csk_reset_xmit_timer(sk, ICSK_TIME_PROBE0, min(icsk->icsk_rto << icsk->icsk_backoff, TCP_RTO_MAX), TCP_RTO_MAX); } else { @@ -2509,7 +2520,7 @@ void tcp_send_probe0(struct sock *sk) */ if (!icsk->icsk_probes_out) icsk->icsk_probes_out = 1; - inet_csk_reset_xmit_timer(sk, ICSK_TIME_PROBE0, + inet_csk_reset_xmit_timer(sk, ICSK_TIME_PROBE0, min(icsk->icsk_rto << icsk->icsk_backoff, TCP_RESOURCE_PROBE_INTERVAL), TCP_RTO_MAX); |