tcp_input.c

来自「基于组件方式开发操作系统的OSKIT源代码」· C语言 代码 · 共 2,144 行 · 第 1/4 页

/*
 * Copyright (c) 1982, 1986, 1988, 1990, 1993, 1994
 *	The Regents of the University of California.  All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. All advertising materials mentioning features or use of this software
 *    must display the following acknowledgement:
 *	This product includes software developed by the University of
 *	California, Berkeley and its contributors.
 * 4. Neither the name of the University nor the names of its contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 *
 *	From: @(#)tcp_input.c	8.5 (Berkeley) 4/10/94
 */

#ifndef TUBA_INCLUDE
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/malloc.h>
#include <sys/mbuf.h>
#include <sys/protosw.h>
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/errno.h>
#include <sys/queue.h>
#include <sys/kernel.h>

#include <net/if.h>
#include <net/route.h>

#include <netinet/in.h>
#include <netinet/in_systm.h>
#include <netinet/ip.h>
#include <netinet/in_pcb.h>
#include <netinet/ip_var.h>
#include <netinet/tcp.h>
#include <netinet/tcp_fsm.h>
#include <netinet/tcp_seq.h>
#include <netinet/tcp_timer.h>
#include <netinet/tcp_var.h>
#include <netinet/tcpip.h>
#ifdef TCPDEBUG
#include <netinet/tcp_debug.h>
struct	tcpiphdr tcp_saveti;
#endif

int	tcprexmtthresh = 3;
tcp_seq	tcp_iss;
tcp_cc	tcp_ccgen;
struct	tcpstat tcpstat;
u_long	tcp_now;
struct inpcbhead tcb;
struct inpcbinfo tcbinfo;

#endif /* TUBA_INCLUDE */

/*
 * Insert segment ti into reassembly queue of tcp with
 * control block tp.  Return TH_FIN if reassembly now includes
 * a segment with FIN.  The macro form does the common case inline
 * (segment is the next to be received on an established connection,
 * and the queue is empty), avoiding linkage into and removal
 * from the queue and repetition of various conversions.
 * Set DELACK for segments received in order, but ack immediately
 * when segments are out of order (so fast retransmit can work).
 */
#ifdef TCP_ACK_HACK
#define	TCP_REASS(tp, ti, m, so, flags) { \
	if ((ti)->ti_seq == (tp)->rcv_nxt && \
	    (tp)->seg_next == (struct tcpiphdr *)(tp) && \
	    (tp)->t_state == TCPS_ESTABLISHED) { \
		if (ti->ti_flags & TH_PUSH) \
			tp->t_flags |= TF_ACKNOW; \
		else \
			tp->t_flags |= TF_DELACK; \
		(tp)->rcv_nxt += (ti)->ti_len; \
		flags = (ti)->ti_flags & TH_FIN; \
		tcpstat.tcps_rcvpack++;\
		tcpstat.tcps_rcvbyte += (ti)->ti_len;\
		sbappend(&(so)->so_rcv, (m)); \
		sorwakeup(so); \
	} else { \
		(flags) = tcp_reass((tp), (ti), (m)); \
		tp->t_flags |= TF_ACKNOW; \
	} \
}
#else
#define	TCP_REASS(tp, ti, m, so, flags) { \
	if ((ti)->ti_seq == (tp)->rcv_nxt && \
	    (tp)->seg_next == (struct tcpiphdr *)(tp) && \
	    (tp)->t_state == TCPS_ESTABLISHED) { \
		tp->t_flags |= TF_DELACK; \
		(tp)->rcv_nxt += (ti)->ti_len; \
		flags = (ti)->ti_flags & TH_FIN; \
		tcpstat.tcps_rcvpack++;\
		tcpstat.tcps_rcvbyte += (ti)->ti_len;\
		sbappend(&(so)->so_rcv, (m)); \
		sorwakeup(so); \
	} else { \
		(flags) = tcp_reass((tp), (ti), (m)); \
		tp->t_flags |= TF_ACKNOW; \
	} \
}
#endif
#ifndef TUBA_INCLUDE

int
tcp_reass(tp, ti, m)
	register struct tcpcb *tp;
	register struct tcpiphdr *ti;
	struct mbuf *m;
{
	register struct tcpiphdr *q;
	struct socket *so = tp->t_inpcb->inp_socket;
	int flags;

	/*
	 * Call with ti==0 after become established to
	 * force pre-ESTABLISHED data up to user socket.
	 */
	if (ti == 0)
		goto present;

	/*
	 * Find a segment which begins after this one does.
	 */
	for (q = tp->seg_next; q != (struct tcpiphdr *)tp;
	    q = (struct tcpiphdr *)q->ti_next)
		if (SEQ_GT(q->ti_seq, ti->ti_seq))
			break;

	/*
	 * If there is a preceding segment, it may provide some of
	 * our data already.  If so, drop the data from the incoming
	 * segment.  If it provides all of our data, drop us.
	 */
	if ((struct tcpiphdr *)q->ti_prev != (struct tcpiphdr *)tp) {
		register int i;
		q = (struct tcpiphdr *)q->ti_prev;
		/* conversion to int (in i) handles seq wraparound */
		i = q->ti_seq + q->ti_len - ti->ti_seq;
		if (i > 0) {
			if (i >= ti->ti_len) {
				tcpstat.tcps_rcvduppack++;
				tcpstat.tcps_rcvdupbyte += ti->ti_len;
				m_freem(m);
				/*
				 * Try to present any queued data
				 * at the left window edge to the user.
				 * This is needed after the 3-WHS
				 * completes.
				 */
				goto present;	/* ??? */
			}
			m_adj(m, i);
			ti->ti_len -= i;
			ti->ti_seq += i;
		}
		q = (struct tcpiphdr *)(q->ti_next);
	}
	tcpstat.tcps_rcvoopack++;
	tcpstat.tcps_rcvoobyte += ti->ti_len;
	REASS_MBUF(ti) = m;		/* XXX */

	/*
	 * While we overlap succeeding segments trim them or,
	 * if they are completely covered, dequeue them.
	 */
	while (q != (struct tcpiphdr *)tp) {
		register int i = (ti->ti_seq + ti->ti_len) - q->ti_seq;
		if (i <= 0)
			break;
		if (i < q->ti_len) {
			q->ti_seq += i;
			q->ti_len -= i;
			m_adj(REASS_MBUF(q), i);
			break;
		}
		q = (struct tcpiphdr *)q->ti_next;
		m = REASS_MBUF((struct tcpiphdr *)q->ti_prev);
		remque(q->ti_prev);
		m_freem(m);
	}

	/*
	 * Stick new segment in its place.
	 */
	insque(ti, q->ti_prev);

present:
	/*
	 * Present data to user, advancing rcv_nxt through
	 * completed sequence space.
	 */
	if (!TCPS_HAVEESTABLISHED(tp->t_state))
		return (0);
	ti = tp->seg_next;
	if (ti == (struct tcpiphdr *)tp || ti->ti_seq != tp->rcv_nxt)
		return (0);
	do {
		tp->rcv_nxt += ti->ti_len;
		flags = ti->ti_flags & TH_FIN;
		remque(ti);
		m = REASS_MBUF(ti);
		ti = (struct tcpiphdr *)ti->ti_next;
		if (so->so_state & SS_CANTRCVMORE)
			m_freem(m);
		else
			sbappend(&so->so_rcv, m);
	} while (ti != (struct tcpiphdr *)tp && ti->ti_seq == tp->rcv_nxt);
	sorwakeup(so);
	return (flags);
}

/*
 * TCP input routine, follows pages 65-76 of the
 * protocol specification dated September, 1981 very closely.
 */
void
tcp_input(m, iphlen)
	register struct mbuf *m;
	int iphlen;
{
	register struct tcpiphdr *ti;
	register struct inpcb *inp;
	caddr_t optp = NULL;
	int optlen = 0;
	int len, tlen, off;
	register struct tcpcb *tp = 0;
	register int tiflags;
	struct socket *so = 0;
	int todrop, acked, ourfinisacked, needoutput = 0;
	struct in_addr laddr;
	int dropsocket = 0;
	int iss = 0;
	u_long tiwin;
	struct tcpopt to;		/* options in this segment */
	struct rmxp_tao *taop;		/* pointer to our TAO cache entry */
	struct rmxp_tao	tao_noncached;	/* in case there's no cached entry */
#ifdef TCPDEBUG
	short ostate = 0;
#endif

	bzero((char *)&to, sizeof(to));

	tcpstat.tcps_rcvtotal++;
	/*
	 * Get IP and TCP header together in first mbuf.
	 * Note: IP leaves IP header in first mbuf.
	 */
	ti = mtod(m, struct tcpiphdr *);
	if (iphlen > sizeof (struct ip))
		ip_stripoptions(m, (struct mbuf *)0);
	if (m->m_len < sizeof (struct tcpiphdr)) {
		if ((m = m_pullup(m, sizeof (struct tcpiphdr))) == 0) {
			tcpstat.tcps_rcvshort++;
			return;
		}
		ti = mtod(m, struct tcpiphdr *);
	}

	/*
	 * Checksum extended TCP header and data.
	 */
	tlen = ((struct ip *)ti)->ip_len;
	len = sizeof (struct ip) + tlen;
	ti->ti_next = ti->ti_prev = 0;
	ti->ti_x1 = 0;
	ti->ti_len = (u_short)tlen;
	HTONS(ti->ti_len);
	ti->ti_sum = in_cksum(m, len);
	if (ti->ti_sum) {
		tcpstat.tcps_rcvbadsum++;
		goto drop;
	}
#endif /* TUBA_INCLUDE */

	/*
	 * Check that TCP offset makes sense,
	 * pull out TCP options and adjust length.		XXX
	 */
	off = ti->ti_off << 2;
	if (off < sizeof (struct tcphdr) || off > tlen) {
		tcpstat.tcps_rcvbadoff++;
		goto drop;
	}
	tlen -= off;
	ti->ti_len = tlen;
	if (off > sizeof (struct tcphdr)) {
		if (m->m_len < sizeof(struct ip) + off) {
			if ((m = m_pullup(m, sizeof (struct ip) + off)) == 0) {
				tcpstat.tcps_rcvshort++;
				return;
			}
			ti = mtod(m, struct tcpiphdr *);
		}
		optlen = off - sizeof (struct tcphdr);
		optp = mtod(m, caddr_t) + sizeof (struct tcpiphdr);
	}
	tiflags = ti->ti_flags;

	/*
	 * Convert TCP protocol specific fields to host format.
	 */
	NTOHL(ti->ti_seq);
	NTOHL(ti->ti_ack);
	NTOHS(ti->ti_win);
	NTOHS(ti->ti_urp);

	/*
	 * 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);

	/*
	 * Locate pcb for segment.
	 */
findpcb:
	/*
	 * First look for an exact match.
	 */
	inp = in_pcblookuphash(&tcbinfo, ti->ti_src, ti->ti_sport,
	    ti->ti_dst, ti->ti_dport);
	/*
	 * ...and if that fails, do a wildcard search.
	 */
	if (inp == NULL) {
		inp = in_pcblookup(&tcb, ti->ti_src, ti->ti_sport,
		    ti->ti_dst, ti->ti_dport, INPLOOKUP_WILDCARD);
	}

	/*
	 * If the state is CLOSED (i.e., TCB does not exist) then
	 * all data in the incoming segment is discarded.
	 * If the TCB exists but is in CLOSED state, it is embryonic,
	 * but should either do a listen or a connect soon.
	 */
	if (inp == NULL)
		goto dropwithreset;
	tp = intotcpcb(inp);
	if (tp == 0)
		goto dropwithreset;
	if (tp->t_state == TCPS_CLOSED)
		goto drop;

	/* Unscale the window into a 32-bit value. */
	if ((tiflags & TH_SYN) == 0)
		tiwin = ti->ti_win << tp->snd_scale;
	else
		tiwin = ti->ti_win;

	so = inp->inp_socket;
	if (so->so_options & (SO_DEBUG|SO_ACCEPTCONN)) {
#ifdef TCPDEBUG
		if (so->so_options & SO_DEBUG) {
			ostate = tp->t_state;
			tcp_saveti = *ti;
		}
#endif
		if (so->so_options & SO_ACCEPTCONN) {
			register struct tcpcb *tp0 = tp;
			struct socket *so2;
			if ((tiflags & (TH_RST|TH_ACK|TH_SYN)) != TH_SYN) {
				/*
				 * Note: dropwithreset makes sure we don't
				 * send a RST in response to a RST.
				 */
				if (tiflags & TH_ACK) {
					tcpstat.tcps_badsyn++;
					goto dropwithreset;
				}
				goto drop;
			}
			so2 = sonewconn(so, 0);
			if (so2 == 0) {
				unsigned int i, j, qlen;

				static int rnd;
				static long old_mono_secs;
				static unsigned int cur_cnt, old_cnt;

				tcpstat.tcps_listendrop++;

				/*
				 * Keep a decaying average of the number
				 * of overruns we've been getting.
				 */
				if ((i = (mono_time.tv_sec -
					  old_mono_secs)) != 0) {
					old_mono_secs = mono_time.tv_sec;
					old_cnt = cur_cnt / i;
					cur_cnt = 0;
				}

				so2 = so->so_q0;
				if (so2 == 0)
					goto drop;

				/*
				 * If we've been getting a lot of hits,
				 * random drop an incomplete connection
				 * from the queue, otherwise, fall through
				 * so we head-drop from the queue.
				 */
				qlen = so->so_q0len;
				if (++cur_cnt > qlen || old_cnt > qlen) {
					rnd = (314159 * rnd + 66329) & 0xffff;
					j = ((qlen + 1) * rnd) >> 16;

					while (j-- && so2)
						so2 = so2->so_q0;
				}
				if (so2) {
					tcp_drop(sototcpcb(so2), ETIMEDOUT);
					so2 = sonewconn(so, 0);
				}
				if (!so2)
					goto drop;
			}
			so = so2;
			/*
			 * This is ugly, but ....
			 *
			 * Mark socket as temporary until we're
			 * committed to keeping it.  The code at
			 * ``drop'' and ``dropwithreset'' check the
			 * flag dropsocket to see if the temporary
			 * socket created here should be discarded.
			 * We mark the socket as discardable until
			 * we're committed to it below in TCPS_LISTEN.
			 */
			dropsocket++;
			inp = (struct inpcb *)so->so_pcb;
			inp->inp_laddr = ti->ti_dst;
			inp->inp_lport = ti->ti_dport;
			in_pcbrehash(inp);
#if BSD>=43
			inp->inp_options = ip_srcroute();
#endif
			tp = intotcpcb(inp);
			tp->t_state = TCPS_LISTEN;
			tp->t_flags |= tp0->t_flags & (TF_NOPUSH|TF_NOOPT);

			/* 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;
	if (TCPS_HAVEESTABLISHED(tp->t_state))
		tp->t_timer[TCPT_KEEP] = tcp_keepidle;

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

	/*
	 * 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.
	 * Make sure that the hidden state-flags are also off.
	 * Since we check for TCPS_ESTABLISHED above, it can only
	 * be TH_NEEDSYN.
	 */
	if (tp->t_state == TCPS_ESTABLISHED &&
	    (tiflags & (TH_SYN|TH_FIN|TH_RST|TH_URG|TH_ACK)) == TH_ACK &&
	    ((tp->t_flags & (TF_NEEDSYN|TF_NEEDFIN)) == 0) &&
	    ((to.to_flag & TOF_TS) == 0 ||
	     TSTMP_GEQ(to.to_tsval, tp->ts_recent)) &&
	    /*
	     * Using the CC option is compulsory if once started:
	     *   the segment is OK if no T/TCP was negotiated or
	     *   if the segment has a CC option equal to CCrecv
	     */
	    ((tp->t_flags & (TF_REQ_CC|TF_RCVD_CC)) != (TF_REQ_CC|TF_RCVD_CC) ||
	     (to.to_flag & TOF_CC) != 0 && to.to_cc == tp->cc_recv) &&
	    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.
		 * NOTE that the test is modified according to the latest
		 * proposal of the [e-mail] list (Braden 1993/04/26).
		 */
		if ((to.to_flag & TOF_TS) != 0 &&
		   SEQ_LEQ(ti->ti_seq, tp->last_ack_sent)) {
			tp->ts_recent_age = tcp_now;
			tp->ts_recent = to.to_tsval;
		}

		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) {
				/*