tcp_output.c

Linux内核源代码 为压缩文件 是<<Linux内核>>一书中的源代码

			TCP_SKB_CB(skb)->when = tcp_time_stamp;
			if (!tcp_transmit_skb(sk, skb_clone(skb, GFP_KERNEL)))
				update_send_head(sk, tp, skb);
			else
				tcp_check_probe_timer(sk, tp);
		}
	} else {
		/* Socket is locked, keep trying until memory is available. */
		for (;;) {
			skb = alloc_skb(MAX_TCP_HEADER, GFP_KERNEL);
			if (skb)
				break;
			current->policy |= SCHED_YIELD;
			schedule();
		}

		/* Reserve space for headers and prepare control bits. */
		skb_reserve(skb, MAX_TCP_HEADER);
		skb->csum = 0;
		TCP_SKB_CB(skb)->flags = (TCPCB_FLAG_ACK | TCPCB_FLAG_FIN);
		TCP_SKB_CB(skb)->sacked = 0;

		/* FIN eats a sequence byte, write_seq advanced by tcp_send_skb(). */
		TCP_SKB_CB(skb)->seq = tp->write_seq;
		TCP_SKB_CB(skb)->end_seq = TCP_SKB_CB(skb)->seq + 1;
		tcp_send_skb(sk, skb, 0, mss_now);
		__tcp_push_pending_frames(sk, tp, mss_now, 1);
	}
}

/* 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
 */
void tcp_send_active_reset(struct sock *sk, int priority)
{
	struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);
	struct sk_buff *skb;

	/* NOTE: No TCP options attached and we never retransmit this. */
	skb = alloc_skb(MAX_TCP_HEADER, priority);
	if (!skb) {
		NET_INC_STATS(TCPAbortFailed);
		return;
	}

	/* Reserve space for headers and prepare control bits. */
	skb_reserve(skb, MAX_TCP_HEADER);
	skb->csum = 0;
	TCP_SKB_CB(skb)->flags = (TCPCB_FLAG_ACK | TCPCB_FLAG_RST);
	TCP_SKB_CB(skb)->sacked = 0;

	/* Send it off. */
	TCP_SKB_CB(skb)->seq = tcp_acceptable_seq(sk, tp);
	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))
		NET_INC_STATS(TCPAbortFailed);
}

/* WARNING: This routine must only be called when we have already sent
 * a SYN packet that crossed the incoming SYN that caused this routine
 * to get called. If this assumption fails then the initial rcv_wnd
 * and rcv_wscale values will not be correct.
 */
int tcp_send_synack(struct sock *sk)
{
	struct sk_buff* skb;

	skb = skb_peek(&sk->write_queue);
	if (skb == NULL || !(TCP_SKB_CB(skb)->flags&TCPCB_FLAG_SYN)) {
		printk(KERN_DEBUG "tcp_send_synack: wrong queue state\n");
		return -EFAULT;
	}
	if (!(TCP_SKB_CB(skb)->flags&TCPCB_FLAG_ACK)) {
		if (skb_cloned(skb)) {
			struct sk_buff *nskb = skb_copy(skb, GFP_ATOMIC);
			if (nskb == NULL)
				return -ENOMEM;
			__skb_unlink(skb, &sk->write_queue);
			__skb_queue_head(&sk->write_queue, nskb);
			tcp_free_skb(sk, skb);
			tcp_charge_skb(sk, nskb);
			skb = nskb;
		}

		TCP_SKB_CB(skb)->flags |= TCPCB_FLAG_ACK;
		TCP_ECN_send_synack(&sk->tp_pinfo.af_tcp, skb);
	}
	TCP_SKB_CB(skb)->when = tcp_time_stamp;
	return tcp_transmit_skb(sk, skb_clone(skb, GFP_ATOMIC));
}

/*
 * Prepare a SYN-ACK.
 */
struct sk_buff * tcp_make_synack(struct sock *sk, struct dst_entry *dst,
				 struct open_request *req)
{
	struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);
	struct tcphdr *th;
	int tcp_header_size;
	struct sk_buff *skb;

	skb = sock_wmalloc(sk, MAX_TCP_HEADER + 15, 1, GFP_ATOMIC);
	if (skb == NULL)
		return NULL;

	/* Reserve space for headers. */
	skb_reserve(skb, MAX_TCP_HEADER);

	skb->dst = dst_clone(dst);

	tcp_header_size = (sizeof(struct tcphdr) + TCPOLEN_MSS +
			   (req->tstamp_ok ? TCPOLEN_TSTAMP_ALIGNED : 0) +
			   (req->wscale_ok ? TCPOLEN_WSCALE_ALIGNED : 0) +
			   /* SACK_PERM is in the place of NOP NOP of TS */
			   ((req->sack_ok && !req->tstamp_ok) ? TCPOLEN_SACKPERM_ALIGNED : 0));
	skb->h.th = th = (struct tcphdr *) skb_push(skb, tcp_header_size);

	memset(th, 0, sizeof(struct tcphdr));
	th->syn = 1;
	th->ack = 1;
	TCP_ECN_make_synack(req, th);
	th->source = sk->sport;
	th->dest = req->rmt_port;
	TCP_SKB_CB(skb)->seq = req->snt_isn;
	TCP_SKB_CB(skb)->end_seq = TCP_SKB_CB(skb)->seq + 1;
	th->seq = htonl(TCP_SKB_CB(skb)->seq);
	th->ack_seq = htonl(req->rcv_isn + 1);
	if (req->rcv_wnd == 0) { /* ignored for retransmitted syns */
		__u8 rcv_wscale; 
		/* Set this up on the first call only */
		req->window_clamp = tp->window_clamp ? : dst->window;
		/* tcp_full_space because it is guaranteed to be the first packet */
		tcp_select_initial_window(tcp_full_space(sk), 
			dst->advmss - (req->tstamp_ok ? TCPOLEN_TSTAMP_ALIGNED : 0),
			&req->rcv_wnd,
			&req->window_clamp,
			req->wscale_ok,
			&rcv_wscale);
		req->rcv_wscale = rcv_wscale; 
	}

	/* RFC1323: The window in SYN & SYN/ACK segments is never scaled. */
	th->window = htons(req->rcv_wnd);

	TCP_SKB_CB(skb)->when = tcp_time_stamp;
	tcp_syn_build_options((__u32 *)(th + 1), dst->advmss, req->tstamp_ok,
			      req->sack_ok, req->wscale_ok, req->rcv_wscale,
			      TCP_SKB_CB(skb)->when,
			      req->ts_recent);

	skb->csum = 0;
	th->doff = (tcp_header_size >> 2);
	TCP_INC_STATS(TcpOutSegs);
	return skb;
}

int tcp_connect(struct sock *sk, struct sk_buff *buff)
{
	struct dst_entry *dst = __sk_dst_get(sk);
	struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);

	/* Reserve space for headers. */
	skb_reserve(buff, MAX_TCP_HEADER);

	/* We'll fix this up when we get a response from the other end.
	 * See tcp_input.c:tcp_rcv_state_process case TCP_SYN_SENT.
	 */
	tp->tcp_header_len = sizeof(struct tcphdr) +
		(sysctl_tcp_timestamps ? TCPOLEN_TSTAMP_ALIGNED : 0);

	/* If user gave his TCP_MAXSEG, record it to clamp */
	if (tp->user_mss)
		tp->mss_clamp = tp->user_mss;
	tp->max_window = 0;
	tcp_sync_mss(sk, dst->pmtu);

	if (!tp->window_clamp)
		tp->window_clamp = dst->window;
	tp->advmss = dst->advmss;
	tcp_initialize_rcv_mss(sk);

	tcp_select_initial_window(tcp_full_space(sk),
				  tp->advmss - (tp->ts_recent_stamp ? tp->tcp_header_len - sizeof(struct tcphdr) : 0),
				  &tp->rcv_wnd,
				  &tp->window_clamp,
				  sysctl_tcp_window_scaling,
				  &tp->rcv_wscale);

	tp->rcv_ssthresh = tp->rcv_wnd;

	/* Socket identity change complete, no longer
	 * in TCP_CLOSE, so enter ourselves into the
	 * hash tables.
	 */
	tcp_set_state(sk,TCP_SYN_SENT);
	if (tp->af_specific->hash_connecting(sk))
		goto err_out;

	sk->err = 0;
	sk->done = 0;
	tp->snd_wnd = 0;
	tcp_init_wl(tp, tp->write_seq, 0);
	tp->snd_una = tp->write_seq;
	tp->snd_sml = tp->write_seq;
	tp->rcv_nxt = 0;
	tp->rcv_wup = 0;
	tp->copied_seq = 0;

	tp->rto = TCP_TIMEOUT_INIT;
	tp->retransmits = 0;
	tcp_clear_retrans(tp);

	TCP_SKB_CB(buff)->flags = TCPCB_FLAG_SYN;
	TCP_ECN_send_syn(tp, buff);
	TCP_SKB_CB(buff)->sacked = 0;
	buff->csum = 0;
	TCP_SKB_CB(buff)->seq = tp->write_seq++;
	TCP_SKB_CB(buff)->end_seq = tp->write_seq;
	tp->snd_nxt = tp->write_seq;
	tp->pushed_seq = tp->write_seq;

	/* Send it off. */
	TCP_SKB_CB(buff)->when = tcp_time_stamp;
	tp->retrans_stamp = TCP_SKB_CB(buff)->when;
	__skb_queue_tail(&sk->write_queue, buff);
	tcp_charge_skb(sk, buff);
	tp->packets_out++;
	tcp_transmit_skb(sk, skb_clone(buff, GFP_KERNEL));
	TCP_INC_STATS(TcpActiveOpens);

	/* Timer for repeating the SYN until an answer. */
	tcp_reset_xmit_timer(sk, TCP_TIME_RETRANS, tp->rto);
	return 0;

err_out:
	tcp_set_state(sk,TCP_CLOSE);
	kfree_skb(buff);
	return -EADDRNOTAVAIL;
}

/* Send out a delayed ack, the caller does the policy checking
 * to see if we should even be here.  See tcp_input.c:tcp_ack_snd_check()
 * for details.
 */
void tcp_send_delayed_ack(struct sock *sk)
{
	struct tcp_opt *tp = &sk->tp_pinfo.af_tcp;
	int ato = tp->ack.ato;
	unsigned long timeout;

	if (ato > TCP_DELACK_MIN) {
		int max_ato = HZ/2;

		if (tp->ack.pingpong || (tp->ack.pending&TCP_ACK_PUSHED))
			max_ato = TCP_DELACK_MAX;

		/* Slow path, intersegment interval is "high". */

		/* If some rtt estimate is known, use it to bound delayed ack.
		 * Do not use tp->rto here, use results of rtt measurements
		 * directly.
		 */
		if (tp->srtt) {
			int rtt = max(tp->srtt>>3, TCP_DELACK_MIN);

			if (rtt < max_ato)
				max_ato = rtt;
		}

		ato = min(ato, max_ato);
	}

	/* Stay within the limit we were given */
	timeout = jiffies + ato;

	/* Use new timeout only if there wasn't a older one earlier. */
	if (tp->ack.pending&TCP_ACK_TIMER) {
		/* If delack timer was blocked or is about to expire,
		 * send ACK now.
		 */
		if (tp->ack.blocked || time_before_eq(tp->ack.timeout, jiffies+(ato>>2))) {
			tcp_send_ack(sk);
			return;
		}

		if (!time_before(timeout, tp->ack.timeout))
			timeout = tp->ack.timeout;
	}
	tp->ack.pending |= TCP_ACK_SCHED|TCP_ACK_TIMER;
	tp->ack.timeout = timeout;
	if (!mod_timer(&tp->delack_timer, timeout))
		sock_hold(sk);

#ifdef TCP_FORMAL_WINDOW
	/* Explanation. Header prediction path does not handle
	 * case of zero window. If we send ACK immediately, pred_flags
	 * are reset when sending ACK. If rcv_nxt is advanced and
	 * ack is not sent, than delayed ack is scheduled.
	 * Hence, it is the best place to check for zero window.
	 */
	if (tp->pred_flags) {
		if (tcp_receive_window(tp) == 0)
			tp->pred_flags = 0;
	} else {
		if (skb_queue_len(&tp->out_of_order_queue) == 0 &&
		    !tp->urg_data)
			tcp_fast_path_on(tp);
	}
#endif
}

/* This routine sends an ack and also updates the window. */
void tcp_send_ack(struct sock *sk)
{
	/* If we have been reset, we may not send again. */
	if(sk->state != TCP_CLOSE) {
		struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);
		struct sk_buff *buff;

		/* We are not putting this on the write queue, so
		 * tcp_transmit_skb() will set the ownership to this
		 * sock.
		 */
		buff = alloc_skb(MAX_TCP_HEADER, GFP_ATOMIC);
		if (buff == NULL) {
			tcp_schedule_ack(tp);
			tp->ack.ato = TCP_ATO_MIN;
			tcp_reset_xmit_timer(sk, TCP_TIME_DACK, TCP_DELACK_MAX);
			return;
		}

		/* Reserve space for headers and prepare control bits. */
		skb_reserve(buff, MAX_TCP_HEADER);
		buff->csum = 0;
		TCP_SKB_CB(buff)->flags = TCPCB_FLAG_ACK;
		TCP_SKB_CB(buff)->sacked = 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)->when = tcp_time_stamp;
		tcp_transmit_skb(sk, buff);
	}
}

/* This routine sends a packet with an out of date sequence
 * number. It assumes the other end will try to ack it.
 *
 * Question: what should we make while urgent mode?
 * 4.4BSD forces sending single byte of data. We cannot send
 * out of window data, because we have SND.NXT==SND.MAX...
 *
 * Current solution: to send TWO zero-length segments in urgent mode:
 * one is with SEG.SEQ=SND.UNA to deliver urgent pointer, another is
 * out-of-date with SND.UNA-1 to probe window.
 */
static int tcp_xmit_probe_skb(struct sock *sk, int urgent)
{
	struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);
	struct sk_buff *skb;

	/* We don't queue it, tcp_transmit_skb() sets ownership. */
	skb = alloc_skb(MAX_TCP_HEADER, GFP_ATOMIC);
	if (skb == NULL) 
		return -1;

	/* Reserve space for headers and set control bits. */
	skb_reserve(skb, MAX_TCP_HEADER);
	skb->csum = 0;
	TCP_SKB_CB(skb)->flags = TCPCB_FLAG_ACK;
	TCP_SKB_CB(skb)->sacked = urgent;

	/* Use a previous sequence.  This should cause the other
	 * end to send an ack.  Don't queue or clone SKB, just
	 * send it.
	 */
	TCP_SKB_CB(skb)->seq = urgent ? tp->snd_una : tp->snd_una - 1;
	TCP_SKB_CB(skb)->end_seq = TCP_SKB_CB(skb)->seq;
	TCP_SKB_CB(skb)->when = tcp_time_stamp;
	return tcp_transmit_skb(sk, skb);
}

int tcp_write_wakeup(struct sock *sk)
{
	if (sk->state != TCP_CLOSE) {
		struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);
		struct sk_buff *skb;

		if ((skb = tp->send_head) != NULL &&
		    before(TCP_SKB_CB(skb)->seq, tp->snd_una+tp->snd_wnd)) {
			int err;
			int mss = tcp_current_mss(sk);
			int seg_size = tp->snd_una+tp->snd_wnd-TCP_SKB_CB(skb)->seq;

			if (before(tp->pushed_seq, TCP_SKB_CB(skb)->end_seq))
				tp->pushed_seq = TCP_SKB_CB(skb)->end_seq;

			/* We are probing the opening of a window
			 * but the window size is != 0
			 * must have been a result SWS avoidance ( sender )
			 */
			if (seg_size < TCP_SKB_CB(skb)->end_seq - TCP_SKB_CB(skb)->seq ||
			    skb->len > mss) {
				seg_size = min(seg_size, mss);
				TCP_SKB_CB(skb)->flags |= TCPCB_FLAG_PSH;
				if (tcp_fragment(sk, skb, seg_size))
					return -1;
			}
			TCP_SKB_CB(skb)->flags |= TCPCB_FLAG_PSH;
			TCP_SKB_CB(skb)->when = tcp_time_stamp;
			err = tcp_transmit_skb(sk, skb_clone(skb, GFP_ATOMIC));
			if (!err) {
				update_send_head(sk, tp, skb);
			}
			return err;
		} else {
			if (tp->urg_mode &&
			    between(tp->snd_up, tp->snd_una+1, tp->snd_una+0xFFFF))
				tcp_xmit_probe_skb(sk, TCPCB_URG);
			return tcp_xmit_probe_skb(sk, 0);
		}
	}
	return -1;
}

/* A window probe timeout has occurred.  If window is not closed send
 * a partial packet else a zero probe.
 */
void tcp_send_probe0(struct sock *sk)
{
	struct tcp_opt *tp = &(sk->tp_pinfo.af_tcp);
	int err;

	err = tcp_write_wakeup(sk);

	if (tp->packets_out || !tp->send_head) {
		/* Cancel probe timer, if it is not required. */
		tp->probes_out = 0;
		tp->backoff = 0;
		return;
	}

	if (err <= 0) {
		tp->backoff++;
		tp->probes_out++;
		tcp_reset_xmit_timer (sk, TCP_TIME_PROBE0, 
				      min(tp->rto << tp->backoff, TCP_RTO_MAX));
	} else {
		/* If packet was not sent due to local congestion,
		 * do not backoff and do not remember probes_out.
		 * Let local senders to fight for local resources.
		 *
		 * Use accumulated backoff yet.
		 */
		if (!tp->probes_out)
			tp->probes_out=1;
		tcp_reset_xmit_timer (sk, TCP_TIME_PROBE0, 
				      min(tp->rto << tp->backoff, TCP_RESOURCE_PROBE_INTERVAL));
	}
}