tcp_output.c

<B>Digital的Unix操作系统VAX 4.2源码</B>

		tcp_setpersist(tp);
	}

	/*
	 * No reason to send a segment, just return.
	 */
	return (0);

send:
	/*
	 * Grab a header mbuf, attaching a copy of data to
	 * be transmitted, and initialize the header from
	 * the template for sends on this connection.
	 */
	MGET(m, M_DONTWAIT, MT_DATA);

	if (m == NULL)
		return (ENOBUFS);
	m->m_off = MMAXOFF - sizeof (struct tcpiphdr);
	m->m_len = sizeof (struct tcpiphdr);
	if (len) {
		if (tp->t_force && len == 1) {
			TCPSTAT(tcps_sndprobe++);
		}
		else if (SEQ_LT(tp->snd_nxt, tp->snd_max)) {
			TCPSTAT(tcps_sndrexmitpack++);
			TCPSTAT(tcps_sndrexmitbyte += len);
		} else {
			TCPSTAT(tcps_sndpack++);
			TCPSTAT(tcps_sndbyte += len);
		}
		m->m_next = m_copy(so->so_snd.sb_mb, off, len);
		if (m->m_next == 0)
			len = 0;
	} else if (tp->t_flags & TF_ACKNOW) {
		TCPSTAT(tcps_sndacks++);
	}
	else if (flags & (TH_SYN|TH_FIN|TH_RST)) {
		TCPSTAT(tcps_sndctrl++);
		if (flags & TH_RST) {
			TCPSTAT(tcps_sndrst++);
		}
	}
	else if (SEQ_GT(tp->snd_up, tp->snd_una)) {
		TCPSTAT(tcps_sndurg++);
	}
	else {
		TCPSTAT(tcps_sndwinup++);
	}

	ti = mtod(m, struct tcpiphdr *);
	if (tp->t_template == 0)
		panic("tcp_output");
	bcopy((caddr_t)tp->t_template, (caddr_t)ti, sizeof (struct tcpiphdr));

	/*
	 * Fill in fields, remembering maximum advertised
	 * window for use in delaying messages about window sizes.
	 * If resending a FIN, be sure not to use a new sequence number.
	 */
	if (flags & TH_FIN && tp->t_flags & TF_SENTFIN && 
	    tp->snd_nxt == tp->snd_max)
		    tp->snd_nxt--;
	ti->ti_seq = htonl(tp->snd_nxt);
	ti->ti_ack = htonl(tp->rcv_nxt);
	/*
	 * Before ESTABLISHED, force sending of initial options
	 * unless TCP set to not do any options.
	 */
	opt = NULL;
	if (flags & TH_SYN && (tp->t_flags & TF_NOOPT) == 0) {
		u_short mss;

		mss = MIN(so->so_rcv.sb_hiwat / 2, tcp_mss(tp));
		if (mss > IP_MSS - sizeof(struct tcpiphdr)) {
			opt = tcp_initopt;
			optlen = sizeof (tcp_initopt);
			*(u_short *)(opt + 2) = htons(mss);
		}
	}
	if (opt) {
		m0 = m->m_next;
		m->m_next = m_get(M_DONTWAIT, MT_DATA);
		if (m->m_next == 0) {
			(void) m_free(m);
			m_freem(m0);
			return (ENOBUFS);
		}
		m->m_next->m_next = m0;
		m0 = m->m_next;
		m0->m_len = optlen;
		bcopy((caddr_t)opt, mtod(m0, caddr_t), optlen);
		opt = (u_char *)(mtod(m0, caddr_t) + optlen);
		while (m0->m_len & 0x3) {
			*opt++ = TCPOPT_EOL;
			m0->m_len++;
		}
		optlen = m0->m_len;
		ti->ti_off = (sizeof (struct tcphdr) + optlen) >> 2;
	}
	ti->ti_flags = flags;
	/*
	 * Calculate receive window. Don't shrink window,
	 * but avoid silly window syndrome.
	 */
	if (win < so->so_rcv.sb_hiwat / 4 && win < tp->t_maxseg)	
		win = 0;
	if (win < (int)(tp->rcv_adv - tp->rcv_nxt))
		win = (int)(tp->rcv_adv - tp->rcv_nxt);
		ti->ti_win = htons((u_short)win);
	if (SEQ_GT(tp->snd_up, tp->snd_nxt)) {
		ti->ti_urp = htons((u_short)(tp->snd_up - tp->snd_nxt));
		ti->ti_flags |= TH_URG;
	} else
		/*
		 * If no urgent pointer to send, then we pull
		 * the urgent pointer to the left edge of the send window
		 * so that it doesn't drift into the send window on sequence
		 * number wraparound.
		 */
		tp->snd_up = tp->snd_una;		/* drag it along */
	/*
	 * If anything to send and we can send it all, set PUSH.
	 * (This will keep happy those implementations which only
	 * give data to the user when a buffer fills or a PUSH comes in.)
	 */
	if (len && off+len == so->so_snd.sb_cc)
		ti->ti_flags |= TH_PUSH;

	/*
	 * Put TCP length in extended header, and then
	 * checksum extended header and data.
	 */
	if (len + optlen)
		ti->ti_len = htons((u_short)(sizeof(struct tcphdr) + 
			optlen + len));
#ifdef TCPLOOPBACK
	/* Don't checksum if we are staying on this machine */
	if (!loopback)
#endif TCPLOOPBACK
		ti->ti_sum = in_cksum(m, sizeof (struct tcpiphdr) + (int)optlen + len);

	/*
	 * In transmit state, time the transmission and arrange for
	 * the retransmit.  In persist state, just set snd_max.
	 */
	if (tp->t_force == 0 || tp->t_timer[TCPT_PERSIST] == 0) {
		tcp_seq startseq = tp->snd_nxt;

		/*
		 * Advance snd_nxt over sequence space of this segment.
		 */
		if (flags & TH_SYN)
			tp->snd_nxt++;
		if (flags & TH_FIN) {
			tp->snd_nxt++;
			tp->t_flags |= TF_SENTFIN;
		}
		tp->snd_nxt += len;
		if (SEQ_GT(tp->snd_nxt, tp->snd_max)) {
			tp->snd_max = tp->snd_nxt;
			/*
			 * Time this transmission if not a retransmission and
			 * not currently timing anything.
			 */
			if (tp->t_rtt == 0) {
				tp->t_rtt = 1;
				tp->t_rtseq = startseq;
				TCPSTAT(tcps_segstimed++);
			}
		}

		/*
		 * Set retransmit timer if not currently set,
		 * and not doing an ack or a keep-alive probe.
		 * Initial value for retransmit timer is smoothed
		 * round-trip time + 2 * round-trip time variance.
		 * Initialize shift counter which is used for backoff
		 * of retransmit time.
		 */
		if (tp->t_timer[TCPT_REXMT] == 0 &&
		    tp->snd_nxt != tp->snd_una) {
			tp->t_timer[TCPT_REXMT] = tp->t_rxtcur;
			if (tp->t_timer[TCPT_PERSIST]) {
			tp->t_timer[TCPT_PERSIST] = 0;
				tp->t_rxtshift = 0;
		}
		}
	} else
		if (SEQ_GT(tp->snd_nxt + len, tp->snd_max))
			tp->snd_max = tp->snd_nxt + len;

	/*
	 * Trace.
	 */
	if (so->so_options & SO_DEBUG)
		tcp_trace(TA_OUTPUT, tp->t_state, tp, ti, 0);

#ifdef TCPLOOPBACK
	if (loopback) {
		/* Fake IP transit and get to TCP layer fast */
		((struct ip *)ti)->ip_hl = sizeof (struct ip) >> 2;
		((struct ip *)ti)->ip_len = sizeof (struct tcphdr) + optlen + len;
		/*
		 * Data (will be) sent.
		 * If this advertises a larger window than any other segment,
		 * then remember the size of the advertised window.
		 * Drop send for purpose of ACK requirements.
		 */
		if (win > 0 && SEQ_GT(tp->rcv_nxt+win, tp->rcv_adv))
			tp->rcv_adv = tp->rcv_nxt + win;
		tp->t_flags &= ~(TF_ACKNOW|TF_DELACK);
		/*
		 * SMP:
		 * Tcp_input() can be called from the interrupt handler so
		 * it needs to be able to lock.  Since we are calling down
		 * from the user level, unlock here, call tcp_input and let
		 * tcp_input pretend that data was from the network.
		 * Note that we are at splnet coming into tcp_output
		 * and we are not changing ipl here.
		 * The so->head check here comes from the fact that on a soclose,
		 * we can get here with both so and so->head locked.  We need to
		 * release both locks because the first thing tcp_input does is
		 * assert the lk_tcb which will cause a lock position panic if
		 * any sockets are locked.
		 */
		{
		int owner = 0;

		if ( (so->so_head) && (so != so->so_head) && smp_owner(&so->so_head->lk_socket) ){
			smp_unlock(&so->so_head->lk_socket);
			owner = 1;
		}
		smp_unlock(&so->lk_socket);
		tcp_input(m);
		s = splnet();
		if ( (so->so_head) && (so != so->so_head) && owner ){
			SO_LOCK(so->so_head);
			owner = 0;
		}
		SO_LOCK(so);
		}
	} else {
#endif TCPLOOPBACK

		/*
	 	 * Fill in IP length and desired time to live and
	 	 * send to IP level.
	 	 */
		((struct ip *)ti)->ip_len = sizeof (struct tcpiphdr) + optlen + len;
		((struct ip *)ti)->ip_ttl = TCP_TTL;
	error = ip_output(m, tp->t_inpcb->inp_options, &tp->t_inpcb->inp_route,
	    so->so_options & SO_DONTROUTE, so);
		if (error)
			return (error);
	TCPSTAT(tcps_sndtotal++);

		/*
	 	 * Data sent (as far as we can tell).
	 	 * If this advertises a larger window than any other segment,
	 	 * then remember the size of the advertised window.
	 * Any pending ACK has now been sent.
	 	 */
		if (win > 0 && SEQ_GT(tp->rcv_nxt+win, tp->rcv_adv))
			tp->rcv_adv = tp->rcv_nxt + win;
		tp->t_flags &= ~(TF_ACKNOW|TF_DELACK);
#ifdef TCPLOOPBACK
	}
#endif
	if (sendalot)
		goto again;
	return (0);
}

tcp_setpersist(tp)
	register struct tcpcb *tp;
{
	register int t = ((tp->t_srtt >> 2) + tp->t_rttvar) >> 1;

	if (tp->t_timer[TCPT_REXMT])
		panic("tcp_output REXMT");
	/*
	 * Start/restart persistance timer.
	 */
	TCPT_RANGESET(tp->t_timer[TCPT_PERSIST],
	    t * tcp_backoff[tp->t_rxtshift],
	    TCPTV_PERSMIN, TCPTV_PERSMAX);
	if (tp->t_rxtshift < TCP_MAXRXTSHIFT)
	tp->t_rxtshift++;
}