diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 25a89eaa669d..5d083855c111 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -355,6 +355,12 @@ static void tcp_fixup_rcvbuf(struct sock *sk) rcvmem = 2 * SKB_TRUESIZE(mss + MAX_TCP_HEADER) * tcp_default_init_rwnd(mss); + /* Dynamic Right Sizing (DRS) has 2 to 3 RTT latency + * Allow enough cushion so that sender is not limited by our window + */ + if (sysctl_tcp_moderate_rcvbuf) + rcvmem <<= 2; + if (sk->sk_rcvbuf < rcvmem) sk->sk_rcvbuf = min(rcvmem, sysctl_tcp_rmem[2]); } @@ -373,6 +379,8 @@ void tcp_init_buffer_space(struct sock *sk) tcp_fixup_sndbuf(sk); tp->rcvq_space.space = tp->rcv_wnd; + tp->rcvq_space.time = tcp_time_stamp; + tp->rcvq_space.seq = tp->copied_seq; maxwin = tcp_full_space(sk); @@ -512,48 +520,62 @@ 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; + int copied; 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); + /* Number of bytes copied to user in last RTT */ + copied = tp->copied_seq - tp->rcvq_space.seq; + if (copied <= tp->rcvq_space.space) + goto new_measure; - space = max(tp->rcvq_space.space, space); + /* A bit of theory : + * copied = bytes received in previous RTT, our base window + * To cope with packet losses, we need a 2x factor + * To cope with slow start, and sender growing its cwin by 100 % + * every RTT, we need a 4x factor, because the ACK we are sending + * now is for the next RTT, not the current one : + * + */ - if (tp->rcvq_space.space != space) { - int rcvmem; + if (sysctl_tcp_moderate_rcvbuf && + !(sk->sk_userlocks & SOCK_RCVBUF_LOCK)) { + int rcvwin, rcvmem, rcvbuf; - tp->rcvq_space.space = space; + /* minimal window to cope with packet losses, assuming + * steady state. Add some cushion because of small variations. + */ + rcvwin = (copied << 1) + 16 * tp->advmss; - if (sysctl_tcp_moderate_rcvbuf && - !(sk->sk_userlocks & SOCK_RCVBUF_LOCK)) { - int new_clamp = space; + /* If rate increased by 25%, + * assume slow start, rcvwin = 3 * copied + * If rate increased by 50%, + * assume sender can use 2x growth, rcvwin = 4 * copied + */ + if (copied >= + tp->rcvq_space.space + (tp->rcvq_space.space >> 2)) { + if (copied >= + tp->rcvq_space.space + (tp->rcvq_space.space >> 1)) + rcvwin <<= 1; + else + rcvwin += (rcvwin >> 1); + } - /* 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 = SKB_TRUESIZE(tp->advmss + MAX_TCP_HEADER); - 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; + rcvmem = SKB_TRUESIZE(tp->advmss + MAX_TCP_HEADER); + while (tcp_win_from_space(rcvmem) < tp->advmss) + rcvmem += 128; - /* Make the window clamp follow along. */ - tp->window_clamp = new_clamp; - } + rcvbuf = min(rcvwin / tp->advmss * rcvmem, sysctl_tcp_rmem[2]); + if (rcvbuf > sk->sk_rcvbuf) { + sk->sk_rcvbuf = rcvbuf; + + /* Make the window clamp follow along. */ + tp->window_clamp = rcvwin; } } + tp->rcvq_space.space = copied; new_measure: tp->rcvq_space.seq = tp->copied_seq; @@ -5674,8 +5696,8 @@ int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb, tcp_init_congestion_control(sk); tcp_mtup_init(sk); - tcp_init_buffer_space(sk); tp->copied_seq = tp->rcv_nxt; + tcp_init_buffer_space(sk); } smp_mb(); tcp_set_state(sk, TCP_ESTABLISHED);