tcp_subr.c

来自「eCos操作系统源码」· C语言 代码 · 共 1,397 行 · 第 1/3 页

//==========================================================================
//
//      src/sys/netinet/tcp_subr.c
//
//==========================================================================
//####BSDCOPYRIGHTBEGIN####
//
// -------------------------------------------
//
// Portions of this software may have been derived from OpenBSD, 
// FreeBSD or other sources, and are covered by the appropriate
// copyright disclaimers included herein.
//
// Portions created by Red Hat are
// Copyright (C) 2002 Red Hat, Inc. All Rights Reserved.
//
// -------------------------------------------
//
//####BSDCOPYRIGHTEND####
//==========================================================================

/*
 * Copyright (c) 1982, 1986, 1988, 1990, 1993, 1995
 *	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.
 *
 *	@(#)tcp_subr.c	8.2 (Berkeley) 5/24/95
 * $FreeBSD: src/sys/netinet/tcp_subr.c,v 1.73.2.22 2001/08/22 00:59:12 silby Exp $
 */

#include <sys/param.h>
#include <sys/malloc.h>
#include <sys/sysctl.h>
#include <sys/mbuf.h>
#ifdef INET6
#include <sys/domain.h>
#endif
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/protosw.h>

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

#define _IP_VHL
#include <netinet/in.h>
#include <netinet/in_systm.h>
#include <netinet/ip.h>
#ifdef INET6
#include <netinet/ip6.h>
#endif
#include <netinet/in_pcb.h>
#ifdef INET6
#include <netinet6/in6_pcb.h>
#endif
#include <netinet/in_var.h>
#include <netinet/ip_var.h>
#ifdef INET6
#include <netinet6/ip6_var.h>
#endif
#include <netinet/tcp.h>
#include <netinet/tcp_fsm.h>
#include <netinet/tcp_seq.h>
#include <netinet/tcp_timer.h>
#include <netinet/tcp_var.h>
#ifdef INET6
#include <netinet6/tcp6_var.h>
#include <netinet6/ip6protosw.h>
#endif
#include <netinet/tcpip.h>
#ifdef TCPDEBUG
#include <netinet/tcp_debug.h>
#endif

#ifdef IPSEC
#include <netinet6/ipsec.h>
#endif /*IPSEC*/

#include <sys/md5.h>

int 	tcp_mssdflt = TCP_MSS;
SYSCTL_INT(_net_inet_tcp, TCPCTL_MSSDFLT, mssdflt, CTLFLAG_RW, 
    &tcp_mssdflt , 0, "Default TCP Maximum Segment Size");

#ifdef INET6
int	tcp_v6mssdflt = TCP6_MSS;
SYSCTL_INT(_net_inet_tcp, TCPCTL_V6MSSDFLT, v6mssdflt,
	CTLFLAG_RW, &tcp_v6mssdflt , 0,
	"Default TCP Maximum Segment Size for IPv6");
#endif

#if 0
static int 	tcp_rttdflt = TCPTV_SRTTDFLT / PR_SLOWHZ;
SYSCTL_INT(_net_inet_tcp, TCPCTL_RTTDFLT, rttdflt, CTLFLAG_RW, 
    &tcp_rttdflt , 0, "Default maximum TCP Round Trip Time");
#endif

static int	tcp_do_rfc1323 = 1;
SYSCTL_INT(_net_inet_tcp, TCPCTL_DO_RFC1323, rfc1323, CTLFLAG_RW, 
    &tcp_do_rfc1323 , 0, "Enable rfc1323 (high performance TCP) extensions");

static int	tcp_do_rfc1644 = 0;
SYSCTL_INT(_net_inet_tcp, TCPCTL_DO_RFC1644, rfc1644, CTLFLAG_RW, 
    &tcp_do_rfc1644 , 0, "Enable rfc1644 (TTCP) extensions");

static int	tcp_tcbhashsize = 0;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, tcbhashsize, CTLFLAG_RD,
     &tcp_tcbhashsize, 0, "Size of TCP control-block hashtable");

static int	do_tcpdrain = 1;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, do_tcpdrain, CTLFLAG_RW, &do_tcpdrain, 0,
     "Enable tcp_drain routine for extra help when low on mbufs");

SYSCTL_INT(_net_inet_tcp, OID_AUTO, pcbcount, CTLFLAG_RD, 
    &tcbinfo.ipi_count, 0, "Number of active PCBs");

static int	icmp_may_rst = 1;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, icmp_may_rst, CTLFLAG_RW, &icmp_may_rst, 0, 
    "Certain ICMP unreachable messages may abort connections in SYN_SENT");

static int	tcp_strict_rfc1948 = 0;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, strict_rfc1948, CTLFLAG_RW,
    &tcp_strict_rfc1948, 0, "Determines if RFC1948 is followed exactly");

static int	tcp_isn_reseed_interval = 0;
SYSCTL_INT(_net_inet_tcp, OID_AUTO, isn_reseed_interval, CTLFLAG_RW,
    &tcp_isn_reseed_interval, 0, "Seconds between reseeding of ISN secret");
    
static void	tcp_cleartaocache __P((void));
static void	tcp_notify __P((struct inpcb *, int));

/*
 * Target size of TCP PCB hash tables. Must be a power of two.
 *
 * Note that this can be overridden by the kernel environment
 * variable net.inet.tcp.tcbhashsize
 */
#ifndef TCBHASHSIZE
#define TCBHASHSIZE	512
#endif

/*
 * This is the actual shape of what we allocate using the zone
 * allocator.  Doing it this way allows us to protect both structures
 * using the same generation count, and also eliminates the overhead
 * of allocating tcpcbs separately.  By hiding the structure here,
 * we avoid changing most of the rest of the code (although it needs
 * to be changed, eventually, for greater efficiency).
 */
#define	ALIGNMENT	32
#define	ALIGNM1		(ALIGNMENT - 1)
struct	inp_tp {
	union {
		struct	inpcb inp;
		char	align[(sizeof(struct inpcb) + ALIGNM1) & ~ALIGNM1];
	} inp_tp_u;
	struct	tcpcb tcb;
	struct	callout inp_tp_rexmt, inp_tp_persist, inp_tp_keep, inp_tp_2msl;
	struct	callout inp_tp_delack;
};
#undef ALIGNMENT
#undef ALIGNM1

/*
 * Tcp initialization
 */
void
tcp_init()
{
	int hashsize = TCBHASHSIZE;
	
	tcp_ccgen = 1;
	tcp_cleartaocache();

	tcp_delacktime = TCPTV_DELACK;
	tcp_keepinit = TCPTV_KEEP_INIT;
	tcp_keepidle = TCPTV_KEEP_IDLE;
	tcp_keepintvl = TCPTV_KEEPINTVL;
	tcp_maxpersistidle = TCPTV_KEEP_IDLE;
	tcp_msl = TCPTV_MSL;
	tcp_rexmit_min = TCPTV_MIN;

	LIST_INIT(&tcb);
	tcbinfo.listhead = &tcb;
	tcp_tcbhashsize = hashsize;
	tcbinfo.hashbase = hashinit(hashsize, M_PCB, &tcbinfo.hashmask);
	tcbinfo.porthashbase = hashinit(hashsize, M_PCB,
					&tcbinfo.porthashmask);
	tcbinfo.ipi_zone = zinit("tcpcb", sizeof(struct inp_tp), maxsockets,
				 ZONE_INTERRUPT, 0);
#ifdef INET6
#define TCP_MINPROTOHDR (sizeof(struct ip6_hdr) + sizeof(struct tcphdr))
#else /* INET6 */
#define TCP_MINPROTOHDR (sizeof(struct tcpiphdr))
#endif /* INET6 */
	if (max_protohdr < TCP_MINPROTOHDR)
		max_protohdr = TCP_MINPROTOHDR;
	if (max_linkhdr + TCP_MINPROTOHDR > MHLEN)
		panic("tcp_init");
#undef TCP_MINPROTOHDR
}

/*
 * Fill in the IP and TCP headers for an outgoing packet, given the tcpcb.
 * tcp_template used to store this data in mbufs, but we now recopy it out
 * of the tcpcb each time to conserve mbufs.
 */
void
tcp_fillheaders(tp, ip_ptr, tcp_ptr)
	struct tcpcb *tp;
	void *ip_ptr;
	void *tcp_ptr;
{
	struct inpcb *inp = tp->t_inpcb;
	struct tcphdr *tcp_hdr = (struct tcphdr *)tcp_ptr;

#ifdef INET6
	if ((inp->inp_vflag & INP_IPV6) != 0) {
		struct ip6_hdr *ip6;

		ip6 = (struct ip6_hdr *)ip_ptr;
		ip6->ip6_flow = (ip6->ip6_flow & ~IPV6_FLOWINFO_MASK) |
			(inp->in6p_flowinfo & IPV6_FLOWINFO_MASK);
		ip6->ip6_vfc = (ip6->ip6_vfc & ~IPV6_VERSION_MASK) |
			(IPV6_VERSION & IPV6_VERSION_MASK);
		ip6->ip6_nxt = IPPROTO_TCP;
		ip6->ip6_plen = sizeof(struct tcphdr);
		ip6->ip6_src = inp->in6p_laddr;
		ip6->ip6_dst = inp->in6p_faddr;
		tcp_hdr->th_sum = 0;
	} else
#endif
	{
	struct ip *ip = (struct ip *) ip_ptr;

	ip->ip_vhl = IP_VHL_BORING;
	ip->ip_tos = 0;
	ip->ip_len = 0;
	ip->ip_id = 0;
	ip->ip_off = 0;
	ip->ip_ttl = 0;
	ip->ip_sum = 0;
	ip->ip_p = IPPROTO_TCP;
	ip->ip_src = inp->inp_laddr;
	ip->ip_dst = inp->inp_faddr;
	tcp_hdr->th_sum = in_pseudo(ip->ip_src.s_addr, ip->ip_dst.s_addr,
		htons(sizeof(struct tcphdr) + IPPROTO_TCP));
	}

	tcp_hdr->th_sport = inp->inp_lport;
	tcp_hdr->th_dport = inp->inp_fport;
	tcp_hdr->th_seq = 0;
	tcp_hdr->th_ack = 0;
	tcp_hdr->th_x2 = 0;
	tcp_hdr->th_off = 5;
	tcp_hdr->th_flags = 0;
	tcp_hdr->th_win = 0;
	tcp_hdr->th_urp = 0;
}

/*
 * Create template to be used to send tcp packets on a connection.
 * Allocates an mbuf and fills in a skeletal tcp/ip header.  The only
 * use for this function is in keepalives, which use tcp_respond.
 */
struct tcptemp *
tcp_maketemplate(tp)
	struct tcpcb *tp;
{
	struct mbuf *m;
	struct tcptemp *n;

	m = m_get(M_DONTWAIT, MT_HEADER);
	if (m == NULL)
		return (0);
	m->m_len = sizeof(struct tcptemp);
	n = mtod(m, struct tcptemp *);

	tcp_fillheaders(tp, (void *)&n->tt_ipgen, (void *)&n->tt_t);
	return (n);
}

/*
 * Send a single message to the TCP at address specified by
 * the given TCP/IP header.  If m == 0, then we make a copy
 * of the tcpiphdr at ti and send directly to the addressed host.
 * This is used to force keep alive messages out using the TCP
 * template for a connection.  If flags are given then we send
 * a message back to the TCP which originated the * segment ti,
 * and discard the mbuf containing it and any other attached mbufs.
 *
 * In any case the ack and sequence number of the transmitted
 * segment are as specified by the parameters.
 *
 * NOTE: If m != NULL, then ti must point to *inside* the mbuf.
 */
void
tcp_respond(tp, ipgen, th, m, ack, seq, flags)
	struct tcpcb *tp;
	void *ipgen;
	register struct tcphdr *th;
	register struct mbuf *m;
	tcp_seq ack, seq;
	int flags;
{
	register int tlen;
	int win = 0;
	struct route *ro = 0;
	struct route sro;
	struct ip *ip;
	struct tcphdr *nth;
#ifdef INET6
#ifdef NEW_STRUCT_ROUTE
	struct route *ro6 = 0;
	struct route sro6;
#else
	struct route_in6 *ro6 = 0;
	struct route_in6 sro6;
#endif
	struct ip6_hdr *ip6;
	int isipv6;
#endif /* INET6 */
	int ipflags = 0;

#ifdef INET6
	isipv6 = IP_VHL_V(((struct ip *)ipgen)->ip_vhl) == 6;
	ip6 = ipgen;
#endif /* INET6 */
	ip = ipgen;

	if (tp) {
		if (!(flags & TH_RST)) {
			win = sbspace(&tp->t_inpcb->inp_socket->so_rcv);
			if (win > (long)TCP_MAXWIN << tp->rcv_scale)
				win = (long)TCP_MAXWIN << tp->rcv_scale;
		}
#ifdef INET6
		if (isipv6)
			ro6 = &tp->t_inpcb->in6p_route;
		else
#endif /* INET6 */
		ro = &tp->t_inpcb->inp_route;
	} else {
#ifdef INET6
		if (isipv6) {
			ro6 = &sro6;
			bzero(ro6, sizeof *ro6);
		} else
#endif /* INET6 */
	      {
		ro = &sro;
		bzero(ro, sizeof *ro);
	      }
	}
	if (m == 0) {
		m = m_gethdr(M_DONTWAIT, MT_HEADER);
		if (m == NULL)
			return;
		tlen = 0;
		m->m_data += max_linkhdr;
#ifdef INET6
		if (isipv6) {
			bcopy((caddr_t)ip6, mtod(m, caddr_t), 
			      sizeof(struct ip6_hdr));
			ip6 = mtod(m, struct ip6_hdr *);
			nth = (struct tcphdr *)(ip6 + 1);
		} else
#endif /* INET6 */
	      {
		bcopy((caddr_t)ip, mtod(m, caddr_t), sizeof(struct ip));
		ip = mtod(m, struct ip *);
		nth = (struct tcphdr *)(ip + 1);
	      }
		bcopy((caddr_t)th, (caddr_t)nth, sizeof(struct tcphdr));
		flags = TH_ACK;
	} else {
		m_freem(m->m_next);
		m->m_next = 0;
		m->m_data = (caddr_t)ipgen;
		/* m_len is set later */
		tlen = 0;
#define xchg(a,b,type) { type t; t=a; a=b; b=t; }
#ifdef INET6
		if (isipv6) {
			xchg(ip6->ip6_dst, ip6->ip6_src, struct in6_addr);
			nth = (struct tcphdr *)(ip6 + 1);
		} else
#endif /* INET6 */
	      {
		xchg(ip->ip_dst.s_addr, ip->ip_src.s_addr, n_long);
		nth = (struct tcphdr *)(ip + 1);
	      }
		if (th != nth) {
			/*
			 * this is usually a case when an extension header
			 * exists between the IPv6 header and the
			 * TCP header.
			 */
			nth->th_sport = th->th_sport;
			nth->th_dport = th->th_dport;
		}
		xchg(nth->th_dport, nth->th_sport, n_short);
#undef xchg
	}
#ifdef INET6
	if (isipv6) {
		ip6->ip6_plen = htons((u_short)(sizeof (struct tcphdr) +
						tlen));
		tlen += sizeof (struct ip6_hdr) + sizeof (struct tcphdr);
	} else
#endif
      {
	tlen += sizeof (struct tcpiphdr);
	ip->ip_len = tlen;
	ip->ip_ttl = ip_defttl;
      }
	m->m_len = tlen;
	m->m_pkthdr.len = tlen;
	m->m_pkthdr.rcvif = (struct ifnet *) 0;
	nth->th_seq = htonl(seq);
	nth->th_ack = htonl(ack);
	nth->th_x2 = 0;
	nth->th_off = sizeof (struct tcphdr) >> 2;
	nth->th_flags = flags;
	if (tp)
		nth->th_win = htons((u_short) (win >> tp->rcv_scale));
	else
		nth->th_win = htons((u_short)win);
	nth->th_urp = 0;
#ifdef INET6
	if (isipv6) {
		nth->th_sum = 0;
		nth->th_sum = in6_cksum(m, IPPROTO_TCP,
					sizeof(struct ip6_hdr),
					tlen - sizeof(struct ip6_hdr));
		ip6->ip6_hlim = in6_selecthlim(tp ? tp->t_inpcb : NULL,
					       ro6 && ro6->ro_rt ?
					       ro6->ro_rt->rt_ifp :
					       NULL);