tcp_input.c

完整的TCP/IP源代码

                        in_pcbrehash(inp);
#if BSD>=43
			inp->inp_options = ip_srcroute();
#endif
			tp = intotcpcb(inp);
			tp->t_state = TCPS_LISTEN;

			/* Compute proper scaling value from buffer space
			 */
			while (tp->request_r_scale < TCP_MAX_WINSHIFT &&
			   TCP_MAXWIN << tp->request_r_scale < so->so_rcv.sb_hiwat)
				tp->request_r_scale++;
		}
	}

	/*
	 * Segment received on connection.
	 * Reset idle time and keep-alive timer.
	 */
	tp->t_idle = 0;
	tp->t_timer[TCPT_KEEP] = tcp_keepidle;

	/*
	 * Process options if not in LISTEN state,
	 * else do it below (after getting remote address).
	 */
	if (optp && tp->t_state != TCPS_LISTEN)
		tcp_dooptions(tp, optp, optlen, ti,
			&ts_present, &ts_val, &ts_ecr);

	/* 
	 * Header prediction: check for the two common cases
	 * of a uni-directional data xfer.  If the packet has
	 * no control flags, is in-sequence, the window didn't
	 * change and we're not retransmitting, it's a
	 * candidate.  If the length is zero and the ack moved
	 * forward, we're the sender side of the xfer.  Just
	 * free the data acked & wake any higher level process
	 * that was blocked waiting for space.  If the length
	 * is non-zero and the ack didn't move, we're the
	 * receiver side.  If we're getting packets in-order
	 * (the reassembly queue is empty), add the data to
	 * the socket buffer and note that we need a delayed ack.
	 */
	if (tp->t_state == TCPS_ESTABLISHED &&
	    (tiflags & (TH_SYN|TH_FIN|TH_RST|TH_URG|TH_ACK)) == TH_ACK &&
	    (!ts_present || TSTMP_GEQ(ts_val, tp->ts_recent)) &&
	    ti->ti_seq == tp->rcv_nxt &&
	    tiwin && tiwin == tp->snd_wnd &&
	    tp->snd_nxt == tp->snd_max) {

		/* 
		 * If last ACK falls within this segment's sequence numbers,
		 *  record the timestamp.
		 */
		if (ts_present && SEQ_LEQ(ti->ti_seq, tp->last_ack_sent)) {
			tp->ts_recent_age = tcp_now;
			tp->ts_recent = ts_val;
		}

		if (ti->ti_len == 0) {
			if (SEQ_GT(ti->ti_ack, tp->snd_una) &&
			    SEQ_LEQ(ti->ti_ack, tp->snd_max) &&
			    tp->snd_cwnd >= tp->snd_wnd) {
				/*
				 * this is a pure ack for outstanding data.
				 */
				++tcpstat.tcps_predack;
				if (ts_present)
					tcp_xmit_timer(tp, tcp_now-ts_ecr+1);
				else if (tp->t_rtt &&
					    SEQ_GT(ti->ti_ack, tp->t_rtseq))
					tcp_xmit_timer(tp, tp->t_rtt);
				acked = ti->ti_ack - tp->snd_una;
				tcpstat.tcps_rcvackpack++;
				tcpstat.tcps_rcvackbyte += acked;
				sbdrop(&so->so_snd, acked);
				tp->snd_una = ti->ti_ack;
				m_freem(m);

				/*
				 * If all outstanding data are acked, stop
				 * retransmit timer, otherwise restart timer
				 * using current (possibly backed-off) value.
				 * If process is waiting for space,
				 * wakeup/selwakeup/signal.  If data
				 * are ready to send, let tcp_output
				 * decide between more output or persist.
				 */
				if (tp->snd_una == tp->snd_max)
					tp->t_timer[TCPT_REXMT] = 0;
				else if (tp->t_timer[TCPT_PERSIST] == 0)
					tp->t_timer[TCPT_REXMT] = tp->t_rxtcur;

				if ((so->so_snd.sb_want > 0) ||
				    (so->so_snd.sb_sel)      ||
				    (sowakeupHook != NULL))
					sowwakeup(so);
				if (so->so_snd.sb_cc)
					(void) tcp_output(tp);
				return;
			}
		} else if (ti->ti_ack == tp->snd_una &&
		    tp->seg_next == (struct tcpiphdr *)tp &&
		    ti->ti_len <= sbspace(&so->so_rcv)) {
			/*
			 * this is a pure, in-sequence data packet
			 * with nothing on the reassembly queue and
			 * we have enough buffer space to take it.
			 */
			++tcpstat.tcps_preddat;
			tp->rcv_nxt += ti->ti_len;
			tcpstat.tcps_rcvpack++;
			tcpstat.tcps_rcvbyte += ti->ti_len;
			/*
			 * Drop TCP, IP headers and TCP options then add data
			 * to socket buffer.
			 */
			m->m_data += sizeof(struct tcpiphdr)+off-sizeof(struct tcphdr);
			m->m_len -= sizeof(struct tcpiphdr)+off-sizeof(struct tcphdr);
			sbappend(&so->so_rcv, m);
			sorwakeup(so);
			if (ti->ti_flags & TH_PUSH)
				tp->t_flags |= TF_ACKNOW;
			else
				tp->t_flags |= TF_DELACK;
			return;
		}
	}

	/*
	 * Drop TCP, IP headers and TCP options.
	 */
	m->m_data += sizeof(struct tcpiphdr)+off-sizeof(struct tcphdr);
	m->m_len  -= sizeof(struct tcpiphdr)+off-sizeof(struct tcphdr);

	/*
	 * Calculate amount of space in receive window,
	 * and then do TCP input processing.
	 * Receive window is amount of space in rcv queue,
	 * but not less than advertised window.
	 */
	{ int win;

	win = sbspace(&so->so_rcv);
	if (win < 0)
		win = 0;
	tp->rcv_wnd = max(win, (int)(tp->rcv_adv - tp->rcv_nxt));
	}

	switch (tp->t_state) {

	/*
	 * If the state is LISTEN then ignore segment if it contains an RST.
	 * If the segment contains an ACK then it is bad and send a RST.
	 * If it does not contain a SYN then it is not interesting; drop it.
	 * Don't bother responding if the destination was a broadcast.
	 * Otherwise initialize tp->rcv_nxt, and tp->irs, select an initial
	 * tp->iss, and send a segment:
	 *     <SEQ=ISS><ACK=RCV_NXT><CTL=SYN,ACK>
	 * Also initialize tp->snd_nxt to tp->iss+1 and tp->snd_una to tp->iss.
	 * Fill in remote peer address fields if not previously specified.
	 * Enter SYN_RECEIVED state, and process any other fields of this
	 * segment in this state.
	 */
	case TCPS_LISTEN: {
		struct mbuf *am;
		register struct sockaddr_in *sin;

#ifdef already_done
		if (tiflags & TH_RST)
			goto drop;
		if (tiflags & TH_ACK)
			goto dropwithreset;
		if ((tiflags & TH_SYN) == 0)
			goto drop;
#endif
		/*
		 * RFC1122 4.2.3.10, p. 104: discard bcast/mcast SYN
		 * in_broadcast() should never return true on a received
		 * packet with M_BCAST not set.
		 */
		if (m->m_flags & (M_BCAST|M_MCAST) ||
		    IN_MULTICAST(ntohl(ti->ti_dst.s_addr)))
			goto drop;

                /* XXX land bug attack fix */
                if (_landAttackSafe)
                    if ((ti->ti_sport == ti->ti_dport) &&
                        (ti->ti_src.s_addr == ti->ti_dst.s_addr))
                        goto drop;
                
		/* XXX */
		am = mBufClGet(M_DONTWAIT, MT_SONAME, CL_SIZE_128, TRUE);
		if (am == NULL)
			goto drop;
		am->m_len = sizeof (struct sockaddr_in);
		sin = mtod(am, struct sockaddr_in *);
		sin->sin_family = AF_INET;
		sin->sin_len = sizeof(*sin);
		sin->sin_addr = ti->ti_src;
		sin->sin_port = ti->ti_sport;
		bzero((caddr_t)sin->sin_zero, sizeof(sin->sin_zero));
		laddr = inp->inp_laddr;
		if (inp->inp_laddr.s_addr == INADDR_ANY)
			inp->inp_laddr = ti->ti_dst;
		if (in_pcbconnect(inp, am)) {
			inp->inp_laddr = laddr;
			(void) m_free(am);
			goto drop;
		}
		(void) m_free(am);
		tp->t_template = tcp_template(tp);
		if (tp->t_template == 0) {
			tp = tcp_drop(tp, ENOBUFS);
			dropsocket = 0;		/* socket is already gone */
			goto drop;
		}
		if (optp)
			tcp_dooptions(tp, optp, optlen, ti,
				&ts_present, &ts_val, &ts_ecr);
		if (iss)
			tp->iss = iss;
		else
			tp->iss = tcp_iss;
		tcp_iss += TCP_ISSINCR/4;
		tp->irs = ti->ti_seq;
		tcp_sendseqinit(tp);
		tcp_rcvseqinit(tp);
		tp->t_flags |= TF_ACKNOW;
		tp->t_state = TCPS_SYN_RECEIVED;
		tp->t_timer[TCPT_KEEP] = tcp_keepinit;
		dropsocket = 0;		/* committed to socket */
		tcpstat.tcps_accepts++;
		goto trimthenstep6;
		}

	/*
	 * If the state is SYN_SENT:
	 *	if seg contains an ACK, but not for our SYN, drop the input.
	 *	if seg contains a RST, then drop the connection.
	 *	if seg does not contain SYN, then drop it.
	 * Otherwise this is an acceptable SYN segment
	 *	initialize tp->rcv_nxt and tp->irs
	 *	if seg contains ack then advance tp->snd_una
	 *	if SYN has been acked change to ESTABLISHED else SYN_RCVD state
	 *	arrange for segment to be acked (eventually)
	 *	continue processing rest of data/controls, beginning with URG
	 */
	case TCPS_SYN_SENT:
		if ((tiflags & TH_ACK) &&
		    (SEQ_LEQ(ti->ti_ack, tp->iss) ||
		     SEQ_GT(ti->ti_ack, tp->snd_max)))
			goto dropwithreset;
		if (tiflags & TH_RST) {
			if (tiflags & TH_ACK)
				tp = tcp_drop(tp, ECONNREFUSED);
			goto drop;
		}
		if ((tiflags & TH_SYN) == 0)
			goto drop;
		if (tiflags & TH_ACK) {
			tp->snd_una = ti->ti_ack;
			if (SEQ_LT(tp->snd_nxt, tp->snd_una))
				tp->snd_nxt = tp->snd_una;
                        tp->t_timer[TCPT_REXMT] = 0;
		}
		tp->irs = ti->ti_seq;
		tcp_rcvseqinit(tp);
		tp->t_flags |= TF_ACKNOW;
		if (tiflags & TH_ACK ) {
			tcpstat.tcps_connects++;
			soisconnected(so);
			tp->t_state = TCPS_ESTABLISHED;
			/* Do window scaling on this connection? */
			if ((tp->t_flags & (TF_RCVD_SCALE|TF_REQ_SCALE)) ==
				(TF_RCVD_SCALE|TF_REQ_SCALE)) {
				tp->snd_scale = tp->requested_s_scale;
				tp->rcv_scale = tp->request_r_scale;
			}
			(void) tcp_reass(tp, (struct tcpiphdr *)0,
				(struct mbuf *)0);
			/*
			 * if we didn't have to retransmit the SYN,
			 * use its rtt as our initial srtt & rtt var.
			 */
			if (tp->t_rtt)
				tcp_xmit_timer(tp, tp->t_rtt);
		} else
			tp->t_state = TCPS_SYN_RECEIVED;

trimthenstep6:
		/*
		 * Advance ti->ti_seq to correspond to first data byte.
		 * If data, trim to stay within window,
		 * dropping FIN if necessary.
		 */
		ti->ti_seq++;
		if (ti->ti_len > tp->rcv_wnd) {
			todrop = ti->ti_len - tp->rcv_wnd;
			m_adj(m, -todrop);
			ti->ti_len = tp->rcv_wnd;
			tiflags &= ~TH_FIN;
			tcpstat.tcps_rcvpackafterwin++;
			tcpstat.tcps_rcvbyteafterwin += todrop;
		}
		tp->snd_wl1 = ti->ti_seq - 1;
		tp->rcv_up = ti->ti_seq;
		goto step6;
	}

	/*
	 * States other than LISTEN or SYN_SENT.
	 * First check timestamp, if present.
	 * Then check that at least some bytes of segment are within 
	 * receive window.  If segment begins before rcv_nxt,
	 * drop leading data (and SYN); if nothing left, just ack.
	 * 
	 * RFC 1323 PAWS: If we have a timestamp reply on this segment
	 * and it's less than ts_recent, drop it.
	 */
	if (ts_present && (tiflags & TH_RST) == 0 && tp->ts_recent &&
	    TSTMP_LT(ts_val, tp->ts_recent)) {

		/* Check to see if ts_recent is over 24 days old.  */
		if ((int)(tcp_now - tp->ts_recent_age) > TCP_PAWS_IDLE) {
			/*
			 * Invalidate ts_recent.  If this segment updates
			 * ts_recent, the age will be reset later and ts_recent
			 * will get a valid value.  If it does not, setting
			 * ts_recent to zero will at least satisfy the
			 * requirement that zero be placed in the timestamp
			 * echo reply when ts_recent isn't valid.  The
			 * age isn't reset until we get a valid ts_recent
			 * because we don't want out-of-order segments to be
			 * dropped when ts_recent is old.
			 */
			tp->ts_recent = 0;
		} else {
			tcpstat.tcps_rcvduppack++;
			tcpstat.tcps_rcvdupbyte += ti->ti_len;
			tcpstat.tcps_pawsdrop++;
			goto dropafterack;
		}
	}

	todrop = tp->rcv_nxt - ti->ti_seq;
	if (todrop > 0) {
		if (tiflags & TH_SYN) {
			tiflags &= ~TH_SYN;
			ti->ti_seq++;
			if (ti->ti_urp > 1) 
				ti->ti_urp--;
			else
				tiflags &= ~TH_URG;
			todrop--;
		}
#if 0  /* simultaneous open bug  */              
		if (todrop >= ti->ti_len) {
			tcpstat.tcps_rcvduppack++;
			tcpstat.tcps_rcvdupbyte += ti->ti_len;
			/*
			 * If segment is just one to the left of the window,
			 * check two special cases:
			 * 1. Don't toss RST in response to 4.2-style keepalive.
			 * 2. If the only thing to drop is a FIN, we can drop
			 *    it, but check the ACK or we will get into FIN
			 *    wars if our FINs crossed (both CLOSING).
			 * In either case, send ACK to resynchronize,
			 * but keep on processing for RST or ACK.
			 */
			if ((tiflags & TH_FIN && todrop == ti->ti_len + 1)
#ifdef TCP_COMPAT_42
			  || (tiflags & TH_RST && ti->ti_seq == tp->rcv_nxt - 1)
#endif
			   ) {
				todrop = ti->ti_len;
				tiflags &= ~TH_FIN;
			} else {
				/*
				 * Handle the case when a bound socket connects
				 * to itself. Allow packets with a SYN and
				 * an ACK to continue with the processing.
				 */
				if (todrop != 0 || (tiflags & TH_ACK) == 0)
					goto dropafterack;
			}
			tp->t_flags |= TF_ACKNOW;
		} else {
			tcpstat.tcps_rcvpartduppack++;
			tcpstat.tcps_rcvpartdupbyte += todrop;
		}
#else /* XXX  simultaneous open bug fix Steven vol II sec 28.8 fig 28.30 */
                if (todrop > ti->ti_len ||
                    todrop == ti->ti_len && (tiflags & TH_FIN) == 0) {

                    /*
                     * Any valid Fin must be to the left of the window.
                     * At this point the FIN must be a duplicate or
                     * out of sequence; drop it.
                     */
                     tiflags &= ~TH_FIN;

                    /*
                     * Send an ACK to resynchronize and drop any data.
                     * But keep on processing for RST or ACK.
                     */
                    tp->t_flags |= TF_ACKNOW;
                    todrop = ti->ti_len;
                }
                tcpstat.tcps_rcvdupbyte += todrop;
                tcpstat.tcps_rcvduppack++;
#endif /* simultaneous open bug fix */                   
		m_adj(m, todrop);
		ti->ti_seq += todrop;
		ti->ti_len -= todrop;
		if (ti->ti_urp > todrop)
			ti->ti_urp -= todrop;
		else {
			tiflags &= ~TH_URG;
			ti->ti_urp = 0;
		}
	}

	/*
	 * If new data are received on a connection after the
	 * user processes are gone, then RST the other end.
	 */
	if ((so->so_state & SS_NOFDREF) &&
	    tp->t_state > TCPS_CLOSE_WAIT && ti->ti_len) {
		tp = tcp_close(tp);
		tcpstat.tcps_rcvafterclose++;
		goto dropwithreset;
	}

	/*
	 * If segment ends after window, drop trailing data
	 * (and PUSH and FIN); if nothing left, just ACK.
	 */
	todrop = (ti->ti_seq+ti->ti_len) - (tp->rcv_nxt+tp->rcv_wnd);
	if (todrop > 0) {