1994-05-24 10:09:53 +00:00
|
|
|
/*
|
1995-10-03 16:54:17 +00:00
|
|
|
* Copyright (c) 1982, 1986, 1988, 1990, 1993, 1994, 1995
|
1994-05-24 10:09:53 +00:00
|
|
|
* 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.
|
|
|
|
*
|
1995-10-03 16:54:17 +00:00
|
|
|
* @(#)tcp_input.c 8.12 (Berkeley) 5/24/95
|
1999-08-28 01:08:13 +00:00
|
|
|
* $FreeBSD$
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
|
|
|
|
1998-12-03 20:23:21 +00:00
|
|
|
#include "opt_ipfw.h" /* for ipfw_fwd */
|
2000-01-09 19:17:30 +00:00
|
|
|
#include "opt_inet6.h"
|
2000-01-15 14:56:38 +00:00
|
|
|
#include "opt_ipsec.h"
|
1997-09-16 18:36:06 +00:00
|
|
|
#include "opt_tcpdebug.h"
|
1999-09-12 17:22:08 +00:00
|
|
|
#include "opt_tcp_input.h"
|
1997-09-16 18:36:06 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <sys/param.h>
|
|
|
|
#include <sys/systm.h>
|
1995-11-09 20:23:09 +00:00
|
|
|
#include <sys/kernel.h>
|
|
|
|
#include <sys/sysctl.h>
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <sys/malloc.h>
|
|
|
|
#include <sys/mbuf.h>
|
1997-04-27 20:01:29 +00:00
|
|
|
#include <sys/proc.h> /* for proc0 declaration */
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <sys/protosw.h>
|
|
|
|
#include <sys/socket.h>
|
|
|
|
#include <sys/socketvar.h>
|
1996-04-04 10:46:44 +00:00
|
|
|
#include <sys/syslog.h>
|
1994-05-24 10:09:53 +00:00
|
|
|
|
1995-10-03 16:54:17 +00:00
|
|
|
#include <machine/cpu.h> /* before tcp_seq.h, for tcp_random18() */
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <net/if.h>
|
|
|
|
#include <net/route.h>
|
|
|
|
|
|
|
|
#include <netinet/in.h>
|
|
|
|
#include <netinet/in_systm.h>
|
|
|
|
#include <netinet/ip.h>
|
2000-07-04 16:35:15 +00:00
|
|
|
#include <netinet/ip_icmp.h> /* for ICMP_BANDLIM */
|
2000-01-09 19:17:30 +00:00
|
|
|
#include <netinet/in_var.h>
|
2000-07-04 16:35:15 +00:00
|
|
|
#include <netinet/icmp_var.h> /* for ICMP_BANDLIM */
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <netinet/in_pcb.h>
|
|
|
|
#include <netinet/ip_var.h>
|
2000-01-09 19:17:30 +00:00
|
|
|
#ifdef INET6
|
2000-07-04 16:35:15 +00:00
|
|
|
#include <netinet/ip6.h>
|
|
|
|
#include <netinet/icmp6.h>
|
|
|
|
#include <netinet6/nd6.h>
|
2000-01-09 19:17:30 +00:00
|
|
|
#include <netinet6/ip6_var.h>
|
2000-07-04 16:35:15 +00:00
|
|
|
#include <netinet6/in6_pcb.h>
|
2000-01-09 19:17:30 +00:00
|
|
|
#endif
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <netinet/tcp.h>
|
|
|
|
#include <netinet/tcp_fsm.h>
|
|
|
|
#include <netinet/tcp_seq.h>
|
|
|
|
#include <netinet/tcp_timer.h>
|
|
|
|
#include <netinet/tcp_var.h>
|
2000-01-09 19:17:30 +00:00
|
|
|
#ifdef INET6
|
|
|
|
#include <netinet6/tcp6_var.h>
|
|
|
|
#endif
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <netinet/tcpip.h>
|
1994-09-15 10:36:56 +00:00
|
|
|
#ifdef TCPDEBUG
|
1994-05-24 10:09:53 +00:00
|
|
|
#include <netinet/tcp_debug.h>
|
2000-01-09 19:17:30 +00:00
|
|
|
|
|
|
|
u_char tcp_saveipgen[40]; /* the size must be of max ip header, now IPv6 */
|
|
|
|
struct tcphdr tcp_savetcp;
|
|
|
|
#endif /* TCPDEBUG */
|
|
|
|
|
|
|
|
#ifdef IPSEC
|
|
|
|
#include <netinet6/ipsec.h>
|
2000-01-15 14:56:38 +00:00
|
|
|
#ifdef INET6
|
|
|
|
#include <netinet6/ipsec6.h>
|
|
|
|
#endif
|
2000-01-09 19:17:30 +00:00
|
|
|
#include <netkey/key.h>
|
|
|
|
#endif /*IPSEC*/
|
|
|
|
|
2000-03-27 19:14:27 +00:00
|
|
|
#include <machine/in_cksum.h>
|
|
|
|
|
2000-01-09 19:17:30 +00:00
|
|
|
MALLOC_DEFINE(M_TSEGQ, "tseg_qent", "TCP segment queue entry");
|
1994-05-24 10:09:53 +00:00
|
|
|
|
1995-11-14 20:34:56 +00:00
|
|
|
static int tcprexmtthresh = 3;
|
1995-02-14 02:35:19 +00:00
|
|
|
tcp_cc tcp_ccgen;
|
1995-11-14 20:34:56 +00:00
|
|
|
|
1995-02-14 02:35:19 +00:00
|
|
|
struct tcpstat tcpstat;
|
2001-06-23 17:17:59 +00:00
|
|
|
SYSCTL_STRUCT(_net_inet_tcp, TCPCTL_STATS, stats, CTLFLAG_RW,
|
1999-05-03 23:57:32 +00:00
|
|
|
&tcpstat , tcpstat, "TCP statistics (struct tcpstat, netinet/tcp_var.h)");
|
1995-11-14 20:34:56 +00:00
|
|
|
|
1996-04-09 07:01:53 +00:00
|
|
|
static int log_in_vain = 0;
|
1996-04-04 10:46:44 +00:00
|
|
|
SYSCTL_INT(_net_inet_tcp, OID_AUTO, log_in_vain, CTLFLAG_RW,
|
1999-05-03 23:57:32 +00:00
|
|
|
&log_in_vain, 0, "Log all incoming TCP connections");
|
1996-04-04 10:46:44 +00:00
|
|
|
|
1999-08-17 12:17:53 +00:00
|
|
|
static int blackhole = 0;
|
|
|
|
SYSCTL_INT(_net_inet_tcp, OID_AUTO, blackhole, CTLFLAG_RW,
|
|
|
|
&blackhole, 0, "Do not send RST when dropping refused connections");
|
|
|
|
|
1998-02-26 05:25:39 +00:00
|
|
|
int tcp_delack_enabled = 1;
|
1998-04-24 10:08:57 +00:00
|
|
|
SYSCTL_INT(_net_inet_tcp, OID_AUTO, delayed_ack, CTLFLAG_RW,
|
1999-05-03 23:57:32 +00:00
|
|
|
&tcp_delack_enabled, 0,
|
|
|
|
"Delay ACK to try and piggyback it onto a data packet");
|
1998-02-26 05:25:39 +00:00
|
|
|
|
1999-09-14 16:14:05 +00:00
|
|
|
#ifdef TCP_DROP_SYNFIN
|
|
|
|
static int drop_synfin = 0;
|
|
|
|
SYSCTL_INT(_net_inet_tcp, OID_AUTO, drop_synfin, CTLFLAG_RW,
|
|
|
|
&drop_synfin, 0, "Drop TCP packets with SYN+FIN set");
|
|
|
|
#endif
|
|
|
|
|
1995-04-09 01:29:31 +00:00
|
|
|
struct inpcbhead tcb;
|
2000-01-09 19:17:30 +00:00
|
|
|
#define tcb6 tcb /* for KAME src sync over BSD*'s */
|
1995-04-09 01:29:31 +00:00
|
|
|
struct inpcbinfo tcbinfo;
|
1994-05-24 10:09:53 +00:00
|
|
|
|
2001-11-22 04:50:44 +00:00
|
|
|
static void tcp_dooptions __P((struct tcpopt *, u_char *, int, int));
|
1995-11-14 20:34:56 +00:00
|
|
|
static void tcp_pulloutofband __P((struct socket *,
|
2001-11-22 04:50:44 +00:00
|
|
|
struct tcphdr *, struct mbuf *, int));
|
2000-01-09 19:17:30 +00:00
|
|
|
static int tcp_reass __P((struct tcpcb *, struct tcphdr *, int *,
|
|
|
|
struct mbuf *));
|
1995-11-14 20:34:56 +00:00
|
|
|
static void tcp_xmit_timer __P((struct tcpcb *, int));
|
2000-05-06 03:31:09 +00:00
|
|
|
static int tcp_newreno __P((struct tcpcb *, struct tcphdr *));
|
1995-11-14 20:34:56 +00:00
|
|
|
|
2000-01-09 19:17:30 +00:00
|
|
|
/* Neighbor Discovery, Neighbor Unreachability Detection Upper layer hint. */
|
|
|
|
#ifdef INET6
|
|
|
|
#define ND6_HINT(tp) \
|
|
|
|
do { \
|
|
|
|
if ((tp) && (tp)->t_inpcb && \
|
|
|
|
((tp)->t_inpcb->inp_vflag & INP_IPV6) != 0 && \
|
|
|
|
(tp)->t_inpcb->in6p_route.ro_rt) \
|
2000-07-04 16:35:15 +00:00
|
|
|
nd6_nud_hint((tp)->t_inpcb->in6p_route.ro_rt, NULL, 0); \
|
2000-01-09 19:17:30 +00:00
|
|
|
} while (0)
|
|
|
|
#else
|
|
|
|
#define ND6_HINT(tp)
|
|
|
|
#endif
|
1994-05-24 10:09:53 +00:00
|
|
|
|
2001-02-25 15:17:24 +00:00
|
|
|
/*
|
2001-12-02 08:49:29 +00:00
|
|
|
* Indicate whether this ack should be delayed. We can delay the ack if
|
2001-12-13 04:02:09 +00:00
|
|
|
* - delayed acks are enabled and
|
|
|
|
* - there is no delayed ack timer in progress and
|
|
|
|
* - our last ack wasn't a 0-sized window. We never want to delay
|
|
|
|
* the ack that opens up a 0-sized window.
|
2001-02-25 15:17:24 +00:00
|
|
|
*/
|
|
|
|
#define DELAY_ACK(tp) \
|
2001-12-02 08:49:29 +00:00
|
|
|
(tcp_delack_enabled && !callout_pending(tp->tt_delack) && \
|
|
|
|
(tp->t_flags & TF_RXWIN0SENT) == 0)
|
2001-02-25 15:17:24 +00:00
|
|
|
|
1995-11-14 20:34:56 +00:00
|
|
|
static int
|
2000-01-09 19:17:30 +00:00
|
|
|
tcp_reass(tp, th, tlenp, m)
|
1994-05-24 10:09:53 +00:00
|
|
|
register struct tcpcb *tp;
|
2000-01-09 19:17:30 +00:00
|
|
|
register struct tcphdr *th;
|
|
|
|
int *tlenp;
|
1994-05-24 10:09:53 +00:00
|
|
|
struct mbuf *m;
|
|
|
|
{
|
2000-01-09 19:17:30 +00:00
|
|
|
struct tseg_qent *q;
|
|
|
|
struct tseg_qent *p = NULL;
|
|
|
|
struct tseg_qent *nq;
|
|
|
|
struct tseg_qent *te;
|
1994-05-24 10:09:53 +00:00
|
|
|
struct socket *so = tp->t_inpcb->inp_socket;
|
|
|
|
int flags;
|
|
|
|
|
|
|
|
/*
|
2000-01-09 19:17:30 +00:00
|
|
|
* Call with th==0 after become established to
|
1994-05-24 10:09:53 +00:00
|
|
|
* force pre-ESTABLISHED data up to user socket.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if (th == 0)
|
1994-05-24 10:09:53 +00:00
|
|
|
goto present;
|
|
|
|
|
2000-01-09 19:17:30 +00:00
|
|
|
/* Allocate a new queue entry. If we can't, just drop the pkt. XXX */
|
|
|
|
MALLOC(te, struct tseg_qent *, sizeof (struct tseg_qent), M_TSEGQ,
|
|
|
|
M_NOWAIT);
|
|
|
|
if (te == NULL) {
|
|
|
|
tcpstat.tcps_rcvmemdrop++;
|
|
|
|
m_freem(m);
|
|
|
|
return (0);
|
|
|
|
}
|
1998-08-24 07:47:39 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Find a segment which begins after this one does.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
LIST_FOREACH(q, &tp->t_segq, tqe_q) {
|
|
|
|
if (SEQ_GT(q->tqe_th->th_seq, th->th_seq))
|
1994-05-24 10:09:53 +00:00
|
|
|
break;
|
2000-01-09 19:17:30 +00:00
|
|
|
p = q;
|
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* 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.
|
|
|
|
*/
|
1998-08-24 07:47:39 +00:00
|
|
|
if (p != NULL) {
|
1994-05-24 10:09:53 +00:00
|
|
|
register int i;
|
|
|
|
/* conversion to int (in i) handles seq wraparound */
|
2000-01-09 19:17:30 +00:00
|
|
|
i = p->tqe_th->th_seq + p->tqe_len - th->th_seq;
|
1994-05-24 10:09:53 +00:00
|
|
|
if (i > 0) {
|
2000-01-09 19:17:30 +00:00
|
|
|
if (i >= *tlenp) {
|
1994-05-24 10:09:53 +00:00
|
|
|
tcpstat.tcps_rcvduppack++;
|
2000-01-09 19:17:30 +00:00
|
|
|
tcpstat.tcps_rcvdupbyte += *tlenp;
|
1994-05-24 10:09:53 +00:00
|
|
|
m_freem(m);
|
2000-01-09 19:17:30 +00:00
|
|
|
FREE(te, M_TSEGQ);
|
1995-02-09 23:13:27 +00:00
|
|
|
/*
|
|
|
|
* Try to present any queued data
|
|
|
|
* at the left window edge to the user.
|
|
|
|
* This is needed after the 3-WHS
|
|
|
|
* completes.
|
|
|
|
*/
|
|
|
|
goto present; /* ??? */
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
m_adj(m, i);
|
2000-01-09 19:17:30 +00:00
|
|
|
*tlenp -= i;
|
|
|
|
th->th_seq += i;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
tcpstat.tcps_rcvoopack++;
|
2000-01-09 19:17:30 +00:00
|
|
|
tcpstat.tcps_rcvoobyte += *tlenp;
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* While we overlap succeeding segments trim them or,
|
|
|
|
* if they are completely covered, dequeue them.
|
|
|
|
*/
|
1998-08-24 07:47:39 +00:00
|
|
|
while (q) {
|
2000-01-09 19:17:30 +00:00
|
|
|
register int i = (th->th_seq + *tlenp) - q->tqe_th->th_seq;
|
1994-05-24 10:09:53 +00:00
|
|
|
if (i <= 0)
|
|
|
|
break;
|
2000-01-09 19:17:30 +00:00
|
|
|
if (i < q->tqe_len) {
|
|
|
|
q->tqe_th->th_seq += i;
|
|
|
|
q->tqe_len -= i;
|
|
|
|
m_adj(q->tqe_m, i);
|
1994-05-24 10:09:53 +00:00
|
|
|
break;
|
|
|
|
}
|
1998-08-24 07:47:39 +00:00
|
|
|
|
2000-01-09 19:17:30 +00:00
|
|
|
nq = LIST_NEXT(q, tqe_q);
|
|
|
|
LIST_REMOVE(q, tqe_q);
|
|
|
|
m_freem(q->tqe_m);
|
|
|
|
FREE(q, M_TSEGQ);
|
1998-08-24 07:47:39 +00:00
|
|
|
q = nq;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
2000-01-09 19:17:30 +00:00
|
|
|
/* Insert the new segment queue entry into place. */
|
|
|
|
te->tqe_m = m;
|
|
|
|
te->tqe_th = th;
|
|
|
|
te->tqe_len = *tlenp;
|
|
|
|
|
1998-08-24 07:47:39 +00:00
|
|
|
if (p == NULL) {
|
2000-01-09 19:17:30 +00:00
|
|
|
LIST_INSERT_HEAD(&tp->t_segq, te, tqe_q);
|
1998-08-24 07:47:39 +00:00
|
|
|
} else {
|
2000-01-09 19:17:30 +00:00
|
|
|
LIST_INSERT_AFTER(p, te, tqe_q);
|
1998-08-24 07:47:39 +00:00
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
present:
|
|
|
|
/*
|
|
|
|
* Present data to user, advancing rcv_nxt through
|
|
|
|
* completed sequence space.
|
|
|
|
*/
|
1995-02-09 23:13:27 +00:00
|
|
|
if (!TCPS_HAVEESTABLISHED(tp->t_state))
|
1994-05-24 10:09:53 +00:00
|
|
|
return (0);
|
2000-01-09 19:17:30 +00:00
|
|
|
q = LIST_FIRST(&tp->t_segq);
|
|
|
|
if (!q || q->tqe_th->th_seq != tp->rcv_nxt)
|
1994-05-24 10:09:53 +00:00
|
|
|
return (0);
|
|
|
|
do {
|
2000-01-09 19:17:30 +00:00
|
|
|
tp->rcv_nxt += q->tqe_len;
|
|
|
|
flags = q->tqe_th->th_flags & TH_FIN;
|
|
|
|
nq = LIST_NEXT(q, tqe_q);
|
|
|
|
LIST_REMOVE(q, tqe_q);
|
1994-05-24 10:09:53 +00:00
|
|
|
if (so->so_state & SS_CANTRCVMORE)
|
2000-01-09 19:17:30 +00:00
|
|
|
m_freem(q->tqe_m);
|
1994-05-24 10:09:53 +00:00
|
|
|
else
|
2000-01-09 19:17:30 +00:00
|
|
|
sbappend(&so->so_rcv, q->tqe_m);
|
|
|
|
FREE(q, M_TSEGQ);
|
1998-08-24 07:47:39 +00:00
|
|
|
q = nq;
|
2000-01-09 19:17:30 +00:00
|
|
|
} while (q && q->tqe_th->th_seq == tp->rcv_nxt);
|
|
|
|
ND6_HINT(tp);
|
1994-05-24 10:09:53 +00:00
|
|
|
sorwakeup(so);
|
|
|
|
return (flags);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* TCP input routine, follows pages 65-76 of the
|
|
|
|
* protocol specification dated September, 1981 very closely.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
#ifdef INET6
|
|
|
|
int
|
|
|
|
tcp6_input(mp, offp, proto)
|
|
|
|
struct mbuf **mp;
|
|
|
|
int *offp, proto;
|
|
|
|
{
|
|
|
|
register struct mbuf *m = *mp;
|
2001-06-11 12:39:29 +00:00
|
|
|
struct in6_ifaddr *ia6;
|
2000-01-09 19:17:30 +00:00
|
|
|
|
|
|
|
IP6_EXTHDR_CHECK(m, *offp, sizeof(struct tcphdr), IPPROTO_DONE);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* draft-itojun-ipv6-tcp-to-anycast
|
|
|
|
* better place to put this in?
|
|
|
|
*/
|
2001-06-11 12:39:29 +00:00
|
|
|
ia6 = ip6_getdstifaddr(m);
|
|
|
|
if (ia6 && (ia6->ia6_flags & IN6_IFF_ANYCAST)) {
|
2000-01-09 19:17:30 +00:00
|
|
|
struct ip6_hdr *ip6;
|
|
|
|
|
|
|
|
ip6 = mtod(m, struct ip6_hdr *);
|
|
|
|
icmp6_error(m, ICMP6_DST_UNREACH, ICMP6_DST_UNREACH_ADDR,
|
|
|
|
(caddr_t)&ip6->ip6_dst - (caddr_t)ip6);
|
|
|
|
return IPPROTO_DONE;
|
|
|
|
}
|
|
|
|
|
2001-09-03 20:03:55 +00:00
|
|
|
tcp_input(m, *offp);
|
2000-01-09 19:17:30 +00:00
|
|
|
return IPPROTO_DONE;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
void
|
2001-09-03 20:03:55 +00:00
|
|
|
tcp_input(m, off0)
|
1994-05-24 10:09:53 +00:00
|
|
|
register struct mbuf *m;
|
2001-09-03 20:03:55 +00:00
|
|
|
int off0;
|
1994-05-24 10:09:53 +00:00
|
|
|
{
|
2000-01-09 19:17:30 +00:00
|
|
|
register struct tcphdr *th;
|
|
|
|
register struct ip *ip = NULL;
|
|
|
|
register struct ipovly *ipov;
|
1994-05-24 10:09:53 +00:00
|
|
|
register struct inpcb *inp;
|
1995-10-03 16:54:17 +00:00
|
|
|
u_char *optp = NULL;
|
1994-05-25 09:21:21 +00:00
|
|
|
int optlen = 0;
|
1994-05-24 10:09:53 +00:00
|
|
|
int len, tlen, off;
|
2000-01-09 19:17:30 +00:00
|
|
|
int drop_hdrlen;
|
1994-05-24 10:09:53 +00:00
|
|
|
register struct tcpcb *tp = 0;
|
2000-01-09 19:17:30 +00:00
|
|
|
register int thflags;
|
1994-05-25 09:21:21 +00:00
|
|
|
struct socket *so = 0;
|
1994-05-24 10:09:53 +00:00
|
|
|
int todrop, acked, ourfinisacked, needoutput = 0;
|
|
|
|
int iss = 0;
|
1995-02-09 23:13:27 +00:00
|
|
|
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 */
|
1994-09-15 10:36:56 +00:00
|
|
|
#ifdef TCPDEBUG
|
|
|
|
short ostate = 0;
|
|
|
|
#endif
|
2000-01-09 19:17:30 +00:00
|
|
|
#ifdef INET6
|
|
|
|
struct ip6_hdr *ip6 = NULL;
|
|
|
|
int isipv6;
|
|
|
|
#endif /* INET6 */
|
2001-02-11 07:39:51 +00:00
|
|
|
int rstreason; /* For badport_bandlim accounting purposes */
|
1994-05-24 10:09:53 +00:00
|
|
|
|
2000-01-09 19:17:30 +00:00
|
|
|
#ifdef INET6
|
|
|
|
isipv6 = (mtod(m, struct ip *)->ip_v == 6) ? 1 : 0;
|
|
|
|
#endif
|
1995-02-09 23:13:27 +00:00
|
|
|
bzero((char *)&to, sizeof(to));
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
tcpstat.tcps_rcvtotal++;
|
2000-01-09 19:17:30 +00:00
|
|
|
|
|
|
|
#ifdef INET6
|
|
|
|
if (isipv6) {
|
|
|
|
/* IP6_EXTHDR_CHECK() is already done at tcp6_input() */
|
|
|
|
ip6 = mtod(m, struct ip6_hdr *);
|
|
|
|
tlen = sizeof(*ip6) + ntohs(ip6->ip6_plen) - off0;
|
|
|
|
if (in6_cksum(m, IPPROTO_TCP, off0, tlen)) {
|
|
|
|
tcpstat.tcps_rcvbadsum++;
|
|
|
|
goto drop;
|
|
|
|
}
|
|
|
|
th = (struct tcphdr *)((caddr_t)ip6 + off0);
|
2001-06-11 12:39:29 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Be proactive about unspecified IPv6 address in source.
|
|
|
|
* As we use all-zero to indicate unbounded/unconnected pcb,
|
|
|
|
* unspecified IPv6 address can be used to confuse us.
|
|
|
|
*
|
|
|
|
* Note that packets with unspecified IPv6 destination is
|
|
|
|
* already dropped in ip6_input.
|
|
|
|
*/
|
|
|
|
if (IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_src)) {
|
|
|
|
/* XXX stat */
|
|
|
|
goto drop;
|
|
|
|
}
|
2000-01-09 19:17:30 +00:00
|
|
|
} else
|
|
|
|
#endif /* INET6 */
|
|
|
|
{
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Get IP and TCP header together in first mbuf.
|
|
|
|
* Note: IP leaves IP header in first mbuf.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if (off0 > sizeof (struct ip)) {
|
1994-05-24 10:09:53 +00:00
|
|
|
ip_stripoptions(m, (struct mbuf *)0);
|
2000-01-09 19:17:30 +00:00
|
|
|
off0 = sizeof(struct ip);
|
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
if (m->m_len < sizeof (struct tcpiphdr)) {
|
|
|
|
if ((m = m_pullup(m, sizeof (struct tcpiphdr))) == 0) {
|
|
|
|
tcpstat.tcps_rcvshort++;
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
}
|
2000-01-09 19:17:30 +00:00
|
|
|
ip = mtod(m, struct ip *);
|
|
|
|
ipov = (struct ipovly *)ip;
|
|
|
|
th = (struct tcphdr *)((caddr_t)ip + off0);
|
2000-03-27 19:14:27 +00:00
|
|
|
tlen = ip->ip_len;
|
|
|
|
|
|
|
|
if (m->m_pkthdr.csum_flags & CSUM_DATA_VALID) {
|
|
|
|
if (m->m_pkthdr.csum_flags & CSUM_PSEUDO_HDR)
|
|
|
|
th->th_sum = m->m_pkthdr.csum_data;
|
|
|
|
else
|
|
|
|
th->th_sum = in_pseudo(ip->ip_src.s_addr,
|
|
|
|
ip->ip_dst.s_addr, htonl(m->m_pkthdr.csum_data +
|
|
|
|
ip->ip_len + IPPROTO_TCP));
|
|
|
|
th->th_sum ^= 0xffff;
|
|
|
|
} else {
|
|
|
|
/*
|
|
|
|
* Checksum extended TCP header and data.
|
|
|
|
*/
|
|
|
|
len = sizeof (struct ip) + tlen;
|
|
|
|
bzero(ipov->ih_x1, sizeof(ipov->ih_x1));
|
|
|
|
ipov->ih_len = (u_short)tlen;
|
|
|
|
HTONS(ipov->ih_len);
|
|
|
|
th->th_sum = in_cksum(m, len);
|
|
|
|
}
|
2000-01-09 19:17:30 +00:00
|
|
|
if (th->th_sum) {
|
1994-05-24 10:09:53 +00:00
|
|
|
tcpstat.tcps_rcvbadsum++;
|
|
|
|
goto drop;
|
|
|
|
}
|
2000-01-09 19:17:30 +00:00
|
|
|
#ifdef INET6
|
|
|
|
/* Re-initialization for later version check */
|
|
|
|
ip->ip_v = IPVERSION;
|
|
|
|
#endif
|
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Check that TCP offset makes sense,
|
|
|
|
* pull out TCP options and adjust length. XXX
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
off = th->th_off << 2;
|
1994-05-24 10:09:53 +00:00
|
|
|
if (off < sizeof (struct tcphdr) || off > tlen) {
|
|
|
|
tcpstat.tcps_rcvbadoff++;
|
|
|
|
goto drop;
|
|
|
|
}
|
2000-01-09 19:17:30 +00:00
|
|
|
tlen -= off; /* tlen is used instead of ti->ti_len */
|
1994-05-24 10:09:53 +00:00
|
|
|
if (off > sizeof (struct tcphdr)) {
|
2000-01-09 19:17:30 +00:00
|
|
|
#ifdef INET6
|
|
|
|
if (isipv6) {
|
|
|
|
IP6_EXTHDR_CHECK(m, off0, off, );
|
|
|
|
ip6 = mtod(m, struct ip6_hdr *);
|
|
|
|
th = (struct tcphdr *)((caddr_t)ip6 + off0);
|
|
|
|
} else
|
|
|
|
#endif /* INET6 */
|
|
|
|
{
|
1994-05-24 10:09:53 +00:00
|
|
|
if (m->m_len < sizeof(struct ip) + off) {
|
|
|
|
if ((m = m_pullup(m, sizeof (struct ip) + off)) == 0) {
|
|
|
|
tcpstat.tcps_rcvshort++;
|
|
|
|
return;
|
|
|
|
}
|
2000-01-09 19:17:30 +00:00
|
|
|
ip = mtod(m, struct ip *);
|
|
|
|
ipov = (struct ipovly *)ip;
|
|
|
|
th = (struct tcphdr *)((caddr_t)ip + off0);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
2000-01-09 19:17:30 +00:00
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
optlen = off - sizeof (struct tcphdr);
|
2000-01-09 19:17:30 +00:00
|
|
|
optp = (u_char *)(th + 1);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
2000-01-09 19:17:30 +00:00
|
|
|
thflags = th->th_flags;
|
1994-05-24 10:09:53 +00:00
|
|
|
|
1999-09-12 17:22:08 +00:00
|
|
|
#ifdef TCP_DROP_SYNFIN
|
|
|
|
/*
|
|
|
|
* If the drop_synfin option is enabled, drop all packets with
|
|
|
|
* both the SYN and FIN bits set. This prevents e.g. nmap from
|
|
|
|
* identifying the TCP/IP stack.
|
|
|
|
*
|
2001-01-24 16:25:36 +00:00
|
|
|
* This is a violation of the TCP specification.
|
1999-09-12 17:22:08 +00:00
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if (drop_synfin && (thflags & (TH_SYN|TH_FIN)) == (TH_SYN|TH_FIN))
|
1999-09-12 17:22:08 +00:00
|
|
|
goto drop;
|
|
|
|
#endif
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Convert TCP protocol specific fields to host format.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
NTOHL(th->th_seq);
|
|
|
|
NTOHL(th->th_ack);
|
|
|
|
NTOHS(th->th_win);
|
|
|
|
NTOHS(th->th_urp);
|
1994-05-24 10:09:53 +00:00
|
|
|
|
1995-04-05 10:32:14 +00:00
|
|
|
/*
|
2000-01-09 19:17:30 +00:00
|
|
|
* Delay droping TCP, IP headers, IPv6 ext headers, and TCP options,
|
|
|
|
* until after ip6_savecontrol() is called and before other functions
|
|
|
|
* which don't want those proto headers.
|
|
|
|
* Because ip6_savecontrol() is going to parse the mbuf to
|
|
|
|
* search for data to be passed up to user-land, it wants mbuf
|
|
|
|
* parameters to be unchanged.
|
1995-04-05 10:32:14 +00:00
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
drop_hdrlen = off0 + off;
|
1995-04-05 10:32:14 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Locate pcb for segment.
|
|
|
|
*/
|
|
|
|
findpcb:
|
1998-07-06 03:20:19 +00:00
|
|
|
#ifdef IPFIREWALL_FORWARD
|
2000-01-09 19:17:30 +00:00
|
|
|
if (ip_fw_fwd_addr != NULL
|
|
|
|
#ifdef INET6
|
|
|
|
&& isipv6 == NULL /* IPv6 support is not yet */
|
|
|
|
#endif /* INET6 */
|
|
|
|
) {
|
1998-07-06 03:20:19 +00:00
|
|
|
/*
|
|
|
|
* Diverted. Pretend to be the destination.
|
|
|
|
* already got one like this?
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
inp = in_pcblookup_hash(&tcbinfo, ip->ip_src, th->th_sport,
|
|
|
|
ip->ip_dst, th->th_dport, 0, m->m_pkthdr.rcvif);
|
1998-07-06 03:20:19 +00:00
|
|
|
if (!inp) {
|
|
|
|
/*
|
|
|
|
* No, then it's new. Try find the ambushing socket
|
|
|
|
*/
|
|
|
|
if (!ip_fw_fwd_addr->sin_port) {
|
2000-01-09 19:17:30 +00:00
|
|
|
inp = in_pcblookup_hash(&tcbinfo, ip->ip_src,
|
|
|
|
th->th_sport, ip_fw_fwd_addr->sin_addr,
|
|
|
|
th->th_dport, 1, m->m_pkthdr.rcvif);
|
1998-07-06 03:20:19 +00:00
|
|
|
} else {
|
|
|
|
inp = in_pcblookup_hash(&tcbinfo,
|
2000-01-09 19:17:30 +00:00
|
|
|
ip->ip_src, th->th_sport,
|
1998-07-06 03:20:19 +00:00
|
|
|
ip_fw_fwd_addr->sin_addr,
|
1999-12-07 17:39:16 +00:00
|
|
|
ntohs(ip_fw_fwd_addr->sin_port), 1,
|
|
|
|
m->m_pkthdr.rcvif);
|
1998-07-06 03:20:19 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
ip_fw_fwd_addr = NULL;
|
|
|
|
} else
|
|
|
|
#endif /* IPFIREWALL_FORWARD */
|
2000-01-09 19:17:30 +00:00
|
|
|
{
|
|
|
|
#ifdef INET6
|
|
|
|
if (isipv6)
|
|
|
|
inp = in6_pcblookup_hash(&tcbinfo, &ip6->ip6_src, th->th_sport,
|
|
|
|
&ip6->ip6_dst, th->th_dport, 1,
|
|
|
|
m->m_pkthdr.rcvif);
|
|
|
|
else
|
|
|
|
#endif /* INET6 */
|
|
|
|
inp = in_pcblookup_hash(&tcbinfo, ip->ip_src, th->th_sport,
|
|
|
|
ip->ip_dst, th->th_dport, 1, m->m_pkthdr.rcvif);
|
|
|
|
}
|
|
|
|
|
|
|
|
#ifdef IPSEC
|
|
|
|
#ifdef INET6
|
|
|
|
if (isipv6) {
|
|
|
|
if (inp != NULL && ipsec6_in_reject_so(m, inp->inp_socket)) {
|
|
|
|
ipsec6stat.in_polvio++;
|
|
|
|
goto drop;
|
|
|
|
}
|
|
|
|
} else
|
|
|
|
#endif /* INET6 */
|
|
|
|
if (inp != NULL && ipsec4_in_reject_so(m, inp->inp_socket)) {
|
|
|
|
ipsecstat.in_polvio++;
|
|
|
|
goto drop;
|
|
|
|
}
|
|
|
|
#endif /*IPSEC*/
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* 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.
|
|
|
|
*/
|
1996-04-04 10:46:44 +00:00
|
|
|
if (inp == NULL) {
|
1999-08-18 15:40:05 +00:00
|
|
|
if (log_in_vain) {
|
2000-01-09 19:17:30 +00:00
|
|
|
#ifdef INET6
|
|
|
|
char dbuf[INET6_ADDRSTRLEN], sbuf[INET6_ADDRSTRLEN];
|
|
|
|
#else /* INET6 */
|
|
|
|
char dbuf[4*sizeof "123"], sbuf[4*sizeof "123"];
|
|
|
|
#endif /* INET6 */
|
|
|
|
|
|
|
|
#ifdef INET6
|
|
|
|
if (isipv6) {
|
|
|
|
strcpy(dbuf, ip6_sprintf(&ip6->ip6_dst));
|
|
|
|
strcpy(sbuf, ip6_sprintf(&ip6->ip6_src));
|
|
|
|
} else
|
|
|
|
#endif
|
|
|
|
{
|
|
|
|
strcpy(dbuf, inet_ntoa(ip->ip_dst));
|
|
|
|
strcpy(sbuf, inet_ntoa(ip->ip_src));
|
|
|
|
}
|
1999-08-18 15:40:05 +00:00
|
|
|
switch (log_in_vain) {
|
|
|
|
case 1:
|
2000-01-09 19:17:30 +00:00
|
|
|
if(thflags & TH_SYN)
|
1999-08-18 15:40:05 +00:00
|
|
|
log(LOG_INFO,
|
|
|
|
"Connection attempt to TCP %s:%d from %s:%d\n",
|
2000-01-09 19:17:30 +00:00
|
|
|
dbuf, ntohs(th->th_dport),
|
|
|
|
sbuf,
|
|
|
|
ntohs(th->th_sport));
|
1999-08-29 21:59:03 +00:00
|
|
|
break;
|
1999-08-18 15:40:05 +00:00
|
|
|
case 2:
|
|
|
|
log(LOG_INFO,
|
|
|
|
"Connection attempt to TCP %s:%d from %s:%d flags:0x%x\n",
|
2000-01-09 19:17:30 +00:00
|
|
|
dbuf, ntohs(th->th_dport), sbuf,
|
|
|
|
ntohs(th->th_sport), thflags);
|
1999-08-29 21:59:03 +00:00
|
|
|
break;
|
1999-08-18 15:40:05 +00:00
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
1996-04-27 18:19:12 +00:00
|
|
|
}
|
1999-08-19 05:22:12 +00:00
|
|
|
if (blackhole) {
|
1999-08-18 15:40:05 +00:00
|
|
|
switch (blackhole) {
|
1999-08-19 05:22:12 +00:00
|
|
|
case 1:
|
2000-01-09 19:17:30 +00:00
|
|
|
if (thflags & TH_SYN)
|
1999-08-19 05:22:12 +00:00
|
|
|
goto drop;
|
|
|
|
break;
|
|
|
|
case 2:
|
|
|
|
goto drop;
|
|
|
|
default:
|
|
|
|
goto drop;
|
1999-08-18 15:40:05 +00:00
|
|
|
}
|
1999-08-19 05:22:12 +00:00
|
|
|
}
|
2001-02-11 07:39:51 +00:00
|
|
|
rstreason = BANDLIM_RST_CLOSEDPORT;
|
|
|
|
goto dropwithreset;
|
1996-04-04 10:46:44 +00:00
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
tp = intotcpcb(inp);
|
2000-12-15 21:45:49 +00:00
|
|
|
if (tp == 0) {
|
2001-02-11 07:39:51 +00:00
|
|
|
rstreason = BANDLIM_RST_CLOSEDPORT;
|
|
|
|
goto dropwithreset;
|
2000-12-15 21:45:49 +00:00
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
if (tp->t_state == TCPS_CLOSED)
|
|
|
|
goto drop;
|
1995-05-30 08:16:23 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/* Unscale the window into a 32-bit value. */
|
2000-01-09 19:17:30 +00:00
|
|
|
if ((thflags & TH_SYN) == 0)
|
|
|
|
tiwin = th->th_win << tp->snd_scale;
|
1994-05-24 10:09:53 +00:00
|
|
|
else
|
2000-01-09 19:17:30 +00:00
|
|
|
tiwin = th->th_win;
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
so = inp->inp_socket;
|
|
|
|
if (so->so_options & (SO_DEBUG|SO_ACCEPTCONN)) {
|
2001-11-22 04:50:44 +00:00
|
|
|
struct in_conninfo inc;
|
1994-09-15 10:36:56 +00:00
|
|
|
#ifdef TCPDEBUG
|
1994-05-24 10:09:53 +00:00
|
|
|
if (so->so_options & SO_DEBUG) {
|
|
|
|
ostate = tp->t_state;
|
2000-01-09 19:17:30 +00:00
|
|
|
#ifdef INET6
|
|
|
|
if (isipv6)
|
|
|
|
bcopy((char *)ip6, (char *)tcp_saveipgen,
|
|
|
|
sizeof(*ip6));
|
|
|
|
else
|
|
|
|
#endif /* INET6 */
|
|
|
|
bcopy((char *)ip, (char *)tcp_saveipgen, sizeof(*ip));
|
|
|
|
tcp_savetcp = *th;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
1994-09-15 10:36:56 +00:00
|
|
|
#endif
|
2001-11-22 04:50:44 +00:00
|
|
|
/* skip if this isn't a listen socket */
|
|
|
|
if ((so->so_options & SO_ACCEPTCONN) == 0)
|
|
|
|
goto after_listen;
|
2000-01-09 19:17:30 +00:00
|
|
|
#ifdef INET6
|
2001-11-22 04:50:44 +00:00
|
|
|
inc.inc_isipv6 = isipv6;
|
|
|
|
if (isipv6) {
|
|
|
|
inc.inc6_faddr = ip6->ip6_src;
|
|
|
|
inc.inc6_laddr = ip6->ip6_dst;
|
|
|
|
inc.inc6_route.ro_rt = NULL; /* XXX */
|
2000-01-09 19:17:30 +00:00
|
|
|
|
2001-11-22 04:50:44 +00:00
|
|
|
} else
|
|
|
|
#endif /* INET6 */
|
|
|
|
{
|
|
|
|
inc.inc_faddr = ip->ip_src;
|
|
|
|
inc.inc_laddr = ip->ip_dst;
|
|
|
|
inc.inc_route.ro_rt = NULL; /* XXX */
|
|
|
|
}
|
|
|
|
inc.inc_fport = th->th_sport;
|
|
|
|
inc.inc_lport = th->th_dport;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If the state is LISTEN then ignore segment if it contains
|
|
|
|
* a 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.
|
|
|
|
*
|
|
|
|
* If the state is SYN_RECEIVED (syncache) and seg contains
|
|
|
|
* an ACK, but not for our SYN/ACK, send a RST. If the seg
|
|
|
|
* contains a RST, check the sequence number to see if it
|
|
|
|
* is a valid reset segment.
|
|
|
|
*/
|
|
|
|
if ((thflags & (TH_RST|TH_ACK|TH_SYN)) != TH_SYN) {
|
|
|
|
if ((thflags & (TH_RST|TH_ACK|TH_SYN)) == TH_ACK) {
|
|
|
|
if (!syncache_expand(&inc, th, &so, m)) {
|
|
|
|
/*
|
|
|
|
* No syncache entry, or ACK was not
|
|
|
|
* for our SYN/ACK. Send a RST.
|
|
|
|
*/
|
1996-10-07 04:32:42 +00:00
|
|
|
tcpstat.tcps_badsyn++;
|
2001-02-11 07:39:51 +00:00
|
|
|
rstreason = BANDLIM_RST_OPENPORT;
|
|
|
|
goto dropwithreset;
|
1996-10-07 04:32:42 +00:00
|
|
|
}
|
2001-11-22 04:50:44 +00:00
|
|
|
if (so == NULL)
|
|
|
|
/*
|
|
|
|
* Could not complete 3-way handshake,
|
|
|
|
* connection is being closed down, and
|
|
|
|
* syncache will free mbuf.
|
|
|
|
*/
|
|
|
|
return;
|
|
|
|
/*
|
|
|
|
* Socket is created in state SYN_RECEIVED.
|
|
|
|
* Continue processing segment.
|
|
|
|
*/
|
|
|
|
inp = sotoinpcb(so);
|
|
|
|
tp = intotcpcb(inp);
|
|
|
|
/*
|
|
|
|
* This is what would have happened in
|
2002-01-04 17:21:27 +00:00
|
|
|
* tcp_output() when the SYN,ACK was sent.
|
2001-11-22 04:50:44 +00:00
|
|
|
*/
|
|
|
|
tp->snd_up = tp->snd_una;
|
|
|
|
tp->snd_max = tp->snd_nxt = tp->iss + 1;
|
|
|
|
tp->last_ack_sent = tp->rcv_nxt;
|
|
|
|
/*
|
|
|
|
* XXX possible bug - it doesn't appear that tp->snd_wnd is unscaled
|
|
|
|
* until the _second_ ACK is received:
|
|
|
|
* rcv SYN (set wscale opts) --> send SYN/ACK, set snd_wnd = window.
|
|
|
|
* rcv ACK, calculate tiwin --> process SYN_RECEIVED, determine wscale,
|
|
|
|
* move to ESTAB, set snd_wnd to tiwin.
|
|
|
|
*/
|
|
|
|
tp->snd_wnd = tiwin; /* unscaled */
|
|
|
|
goto after_listen;
|
|
|
|
}
|
|
|
|
if (thflags & TH_RST) {
|
|
|
|
syncache_chkrst(&inc, th);
|
1996-10-07 04:32:42 +00:00
|
|
|
goto drop;
|
|
|
|
}
|
2001-11-22 04:50:44 +00:00
|
|
|
if (thflags & TH_ACK) {
|
|
|
|
syncache_badack(&inc);
|
|
|
|
tcpstat.tcps_badsyn++;
|
|
|
|
rstreason = BANDLIM_RST_OPENPORT;
|
|
|
|
goto dropwithreset;
|
|
|
|
}
|
|
|
|
goto drop;
|
|
|
|
}
|
2001-06-11 12:39:29 +00:00
|
|
|
|
2001-11-22 04:50:44 +00:00
|
|
|
/*
|
|
|
|
* Segment's flags are (SYN) or (SYN|FIN).
|
|
|
|
*/
|
2001-06-11 12:39:29 +00:00
|
|
|
#ifdef INET6
|
2001-11-22 04:50:44 +00:00
|
|
|
/*
|
|
|
|
* If deprecated address is forbidden,
|
|
|
|
* we do not accept SYN to deprecated interface
|
|
|
|
* address to prevent any new inbound connection from
|
|
|
|
* getting established.
|
|
|
|
* When we do not accept SYN, we send a TCP RST,
|
|
|
|
* with deprecated source address (instead of dropping
|
|
|
|
* it). We compromise it as it is much better for peer
|
|
|
|
* to send a RST, and RST will be the final packet
|
|
|
|
* for the exchange.
|
|
|
|
*
|
|
|
|
* If we do not forbid deprecated addresses, we accept
|
|
|
|
* the SYN packet. RFC2462 does not suggest dropping
|
|
|
|
* SYN in this case.
|
|
|
|
* If we decipher RFC2462 5.5.4, it says like this:
|
|
|
|
* 1. use of deprecated addr with existing
|
|
|
|
* communication is okay - "SHOULD continue to be
|
|
|
|
* used"
|
|
|
|
* 2. use of it with new communication:
|
|
|
|
* (2a) "SHOULD NOT be used if alternate address
|
|
|
|
* with sufficient scope is available"
|
|
|
|
* (2b) nothing mentioned otherwise.
|
|
|
|
* Here we fall into (2b) case as we have no choice in
|
|
|
|
* our source address selection - we must obey the peer.
|
|
|
|
*
|
|
|
|
* The wording in RFC2462 is confusing, and there are
|
|
|
|
* multiple description text for deprecated address
|
|
|
|
* handling - worse, they are not exactly the same.
|
|
|
|
* I believe 5.5.4 is the best one, so we follow 5.5.4.
|
|
|
|
*/
|
|
|
|
if (isipv6 && !ip6_use_deprecated) {
|
|
|
|
struct in6_ifaddr *ia6;
|
2001-06-11 12:39:29 +00:00
|
|
|
|
2001-11-22 04:50:44 +00:00
|
|
|
if ((ia6 = ip6_getdstifaddr(m)) &&
|
|
|
|
(ia6->ia6_flags & IN6_IFF_DEPRECATED)) {
|
|
|
|
tp = NULL;
|
|
|
|
rstreason = BANDLIM_RST_OPENPORT;
|
|
|
|
goto dropwithreset;
|
1996-02-26 21:47:13 +00:00
|
|
|
}
|
2001-11-22 04:50:44 +00:00
|
|
|
}
|
2000-01-09 19:17:30 +00:00
|
|
|
#endif
|
2001-11-22 04:50:44 +00:00
|
|
|
/*
|
|
|
|
* If it is from this socket, drop it, it must be forged.
|
|
|
|
* Don't bother responding if the destination was a broadcast.
|
|
|
|
*/
|
|
|
|
if (th->th_dport == th->th_sport) {
|
2000-01-09 19:17:30 +00:00
|
|
|
#ifdef INET6
|
2001-11-22 04:50:44 +00:00
|
|
|
if (isipv6) {
|
|
|
|
if (IN6_ARE_ADDR_EQUAL(&ip6->ip6_dst,
|
|
|
|
&ip6->ip6_src))
|
|
|
|
goto drop;
|
|
|
|
} else
|
2000-01-09 19:17:30 +00:00
|
|
|
#endif /* INET6 */
|
2001-11-22 04:50:44 +00:00
|
|
|
if (ip->ip_dst.s_addr == ip->ip_src.s_addr)
|
|
|
|
goto drop;
|
|
|
|
}
|
|
|
|
/*
|
|
|
|
* 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.
|
|
|
|
*
|
|
|
|
* Packets with a multicast source address should also
|
|
|
|
* be discarded.
|
|
|
|
*/
|
|
|
|
if (m->m_flags & (M_BCAST|M_MCAST))
|
|
|
|
goto drop;
|
2000-01-09 19:17:30 +00:00
|
|
|
#ifdef INET6
|
2001-11-22 04:50:44 +00:00
|
|
|
if (isipv6) {
|
|
|
|
if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) ||
|
|
|
|
IN6_IS_ADDR_MULTICAST(&ip6->ip6_src))
|
|
|
|
goto drop;
|
|
|
|
} else
|
|
|
|
#endif
|
|
|
|
if (IN_MULTICAST(ntohl(ip->ip_dst.s_addr)) ||
|
|
|
|
IN_MULTICAST(ntohl(ip->ip_src.s_addr)) ||
|
|
|
|
ip->ip_src.s_addr == htonl(INADDR_BROADCAST))
|
|
|
|
goto drop;
|
|
|
|
/*
|
|
|
|
* SYN appears to be valid; create compressed TCP state
|
|
|
|
* for syncache, or perform t/tcp connection.
|
|
|
|
*/
|
|
|
|
if (so->so_qlen <= so->so_qlimit) {
|
|
|
|
tcp_dooptions(&to, optp, optlen, 1);
|
|
|
|
if (!syncache_add(&inc, &to, th, &so, m))
|
|
|
|
goto drop;
|
|
|
|
if (so == NULL)
|
Improved connection establishment performance by doing local port lookups via
a hashed port list. In the new scheme, in_pcblookup() goes away and is
replaced by a new routine, in_pcblookup_local() for doing the local port
check. Note that this implementation is space inefficient in that the PCB
struct is now too large to fit into 128 bytes. I might deal with this in the
future by using the new zone allocator, but I wanted these changes to be
extensively tested in their current form first.
Also:
1) Fixed off-by-one errors in the port lookup loops in in_pcbbind().
2) Got rid of some unneeded rehashing. Adding a new routine, in_pcbinshash()
to do the initialial hash insertion.
3) Renamed in_pcblookuphash() to in_pcblookup_hash() for easier readability.
4) Added a new routine, in_pcbremlists() to remove the PCB from the various
hash lists.
5) Added/deleted comments where appropriate.
6) Removed unnecessary splnet() locking. In general, the PCB functions should
be called at splnet()...there are unfortunately a few exceptions, however.
7) Reorganized a few structs for better cache line behavior.
8) Killed my TCP_ACK_HACK kludge. It may come back in a different form in
the future, however.
These changes have been tested on wcarchive for more than a month. In tests
done here, connection establishment overhead is reduced by more than 50
times, thus getting rid of one of the major networking scalability problems.
Still to do: make tcp_fastimo/tcp_slowtimo scale well for systems with a
large number of connections. tcp_fastimo is easy; tcp_slowtimo is difficult.
WARNING: Anything that knows about inpcb and tcpcb structs will have to be
recompiled; at the very least, this includes netstat(1).
1998-01-27 09:15:13 +00:00
|
|
|
/*
|
2001-11-22 04:50:44 +00:00
|
|
|
* Entry added to syncache, mbuf used to
|
|
|
|
* send SYN,ACK packet.
|
Improved connection establishment performance by doing local port lookups via
a hashed port list. In the new scheme, in_pcblookup() goes away and is
replaced by a new routine, in_pcblookup_local() for doing the local port
check. Note that this implementation is space inefficient in that the PCB
struct is now too large to fit into 128 bytes. I might deal with this in the
future by using the new zone allocator, but I wanted these changes to be
extensively tested in their current form first.
Also:
1) Fixed off-by-one errors in the port lookup loops in in_pcbbind().
2) Got rid of some unneeded rehashing. Adding a new routine, in_pcbinshash()
to do the initialial hash insertion.
3) Renamed in_pcblookuphash() to in_pcblookup_hash() for easier readability.
4) Added a new routine, in_pcbremlists() to remove the PCB from the various
hash lists.
5) Added/deleted comments where appropriate.
6) Removed unnecessary splnet() locking. In general, the PCB functions should
be called at splnet()...there are unfortunately a few exceptions, however.
7) Reorganized a few structs for better cache line behavior.
8) Killed my TCP_ACK_HACK kludge. It may come back in a different form in
the future, however.
These changes have been tested on wcarchive for more than a month. In tests
done here, connection establishment overhead is reduced by more than 50
times, thus getting rid of one of the major networking scalability problems.
Still to do: make tcp_fastimo/tcp_slowtimo scale well for systems with a
large number of connections. tcp_fastimo is easy; tcp_slowtimo is difficult.
WARNING: Anything that knows about inpcb and tcpcb structs will have to be
recompiled; at the very least, this includes netstat(1).
1998-01-27 09:15:13 +00:00
|
|
|
*/
|
2001-11-22 04:50:44 +00:00
|
|
|
return;
|
2000-01-09 19:17:30 +00:00
|
|
|
/*
|
2001-11-22 04:50:44 +00:00
|
|
|
* Segment passed TAO tests.
|
2000-01-09 19:17:30 +00:00
|
|
|
*/
|
2001-11-22 04:50:44 +00:00
|
|
|
inp = sotoinpcb(so);
|
|
|
|
tp = intotcpcb(inp);
|
|
|
|
tp->snd_wnd = tiwin;
|
|
|
|
tp->t_starttime = ticks;
|
|
|
|
tp->t_state = TCPS_ESTABLISHED;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If there is a FIN, or if there is data and the
|
|
|
|
* connection is local, then delay SYN,ACK(SYN) in
|
|
|
|
* the hope of piggy-backing it on a response
|
|
|
|
* segment. Otherwise must send ACK now in case
|
|
|
|
* the other side is slow starting.
|
|
|
|
*/
|
|
|
|
if (DELAY_ACK(tp) && ((thflags & TH_FIN) ||
|
|
|
|
(tlen != 0 &&
|
|
|
|
#ifdef INET6
|
|
|
|
((isipv6 && in6_localaddr(&inp->in6p_faddr))
|
|
|
|
||
|
|
|
|
(!isipv6 &&
|
2000-01-09 19:17:30 +00:00
|
|
|
#endif
|
2001-11-22 04:50:44 +00:00
|
|
|
in_localaddr(inp->inp_faddr)
|
2000-01-09 19:17:30 +00:00
|
|
|
#ifdef INET6
|
2001-11-22 04:50:44 +00:00
|
|
|
))
|
2000-01-09 19:17:30 +00:00
|
|
|
#endif
|
2001-11-22 04:50:44 +00:00
|
|
|
))) {
|
|
|
|
callout_reset(tp->tt_delack, tcp_delacktime,
|
|
|
|
tcp_timer_delack, tp);
|
|
|
|
tp->t_flags |= TF_NEEDSYN;
|
|
|
|
} else
|
|
|
|
tp->t_flags |= (TF_ACKNOW | TF_NEEDSYN);
|
|
|
|
|
|
|
|
tcpstat.tcps_connects++;
|
|
|
|
soisconnected(so);
|
|
|
|
goto trimthenstep6;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
2001-11-22 04:50:44 +00:00
|
|
|
goto drop;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
2001-11-22 04:50:44 +00:00
|
|
|
after_listen:
|
|
|
|
|
|
|
|
/* XXX temp debugging */
|
|
|
|
/* should not happen - syncache should pick up these connections */
|
|
|
|
if (tp->t_state == TCPS_LISTEN)
|
|
|
|
panic("tcp_input: TCPS_LISTEN");
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Segment received on connection.
|
|
|
|
* Reset idle time and keep-alive timer.
|
|
|
|
*/
|
1999-08-30 21:17:07 +00:00
|
|
|
tp->t_rcvtime = ticks;
|
1996-09-13 18:47:03 +00:00
|
|
|
if (TCPS_HAVEESTABLISHED(tp->t_state))
|
1999-08-30 21:17:07 +00:00
|
|
|
callout_reset(tp->tt_keep, tcp_keepidle, tcp_timer_keep, tp);
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
/*
|
2001-11-22 04:50:44 +00:00
|
|
|
* Process options.
|
|
|
|
* XXX this is tradtitional behavior, may need to be cleaned up.
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
2001-11-22 04:50:44 +00:00
|
|
|
tcp_dooptions(&to, optp, optlen, thflags & TH_SYN);
|
|
|
|
if (thflags & TH_SYN) {
|
|
|
|
if (to.to_flags & TOF_SCALE) {
|
|
|
|
tp->t_flags |= TF_RCVD_SCALE;
|
|
|
|
tp->requested_s_scale = to.to_requested_s_scale;
|
|
|
|
}
|
|
|
|
if (to.to_flags & TOF_TS) {
|
|
|
|
tp->t_flags |= TF_RCVD_TSTMP;
|
|
|
|
tp->ts_recent = to.to_tsval;
|
|
|
|
tp->ts_recent_age = ticks;
|
|
|
|
}
|
|
|
|
if (to.to_flags & (TOF_CC|TOF_CCNEW))
|
|
|
|
tp->t_flags |= TF_RCVD_CC;
|
|
|
|
if (to.to_flags & TOF_MSS)
|
|
|
|
tcp_mss(tp, to.to_mss);
|
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
|
1995-05-30 08:16:23 +00:00
|
|
|
/*
|
1994-05-24 10:09:53 +00:00
|
|
|
* 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.
|
1995-02-09 23:13:27 +00:00
|
|
|
* Make sure that the hidden state-flags are also off.
|
|
|
|
* Since we check for TCPS_ESTABLISHED above, it can only
|
|
|
|
* be TH_NEEDSYN.
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
|
|
|
if (tp->t_state == TCPS_ESTABLISHED &&
|
2000-01-09 19:17:30 +00:00
|
|
|
(thflags & (TH_SYN|TH_FIN|TH_RST|TH_URG|TH_ACK)) == TH_ACK &&
|
1995-02-09 23:13:27 +00:00
|
|
|
((tp->t_flags & (TF_NEEDSYN|TF_NEEDFIN)) == 0) &&
|
2001-11-22 04:50:44 +00:00
|
|
|
((to.to_flags & TOF_TS) == 0 ||
|
1995-02-09 23:13:27 +00:00
|
|
|
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) ||
|
2001-11-22 04:50:44 +00:00
|
|
|
((to.to_flags & TOF_CC) != 0 && to.to_cc == tp->cc_recv)) &&
|
2000-01-09 19:17:30 +00:00
|
|
|
th->th_seq == tp->rcv_nxt &&
|
1994-05-24 10:09:53 +00:00
|
|
|
tiwin && tiwin == tp->snd_wnd &&
|
|
|
|
tp->snd_nxt == tp->snd_max) {
|
|
|
|
|
1995-05-30 08:16:23 +00:00
|
|
|
/*
|
1994-05-24 10:09:53 +00:00
|
|
|
* If last ACK falls within this segment's sequence numbers,
|
1995-02-09 23:13:27 +00:00
|
|
|
* record the timestamp.
|
|
|
|
* NOTE that the test is modified according to the latest
|
|
|
|
* proposal of the tcplw@cray.com list (Braden 1993/04/26).
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
2001-11-22 04:50:44 +00:00
|
|
|
if ((to.to_flags & TOF_TS) != 0 &&
|
2000-01-09 19:17:30 +00:00
|
|
|
SEQ_LEQ(th->th_seq, tp->last_ack_sent)) {
|
1999-08-30 21:17:07 +00:00
|
|
|
tp->ts_recent_age = ticks;
|
1995-02-09 23:13:27 +00:00
|
|
|
tp->ts_recent = to.to_tsval;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
2000-01-09 19:17:30 +00:00
|
|
|
if (tlen == 0) {
|
|
|
|
if (SEQ_GT(th->th_ack, tp->snd_una) &&
|
|
|
|
SEQ_LEQ(th->th_ack, tp->snd_max) &&
|
1996-03-22 18:09:21 +00:00
|
|
|
tp->snd_cwnd >= tp->snd_wnd &&
|
|
|
|
tp->t_dupacks < tcprexmtthresh) {
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* this is a pure ack for outstanding data.
|
|
|
|
*/
|
|
|
|
++tcpstat.tcps_predack;
|
1999-08-30 21:17:07 +00:00
|
|
|
/*
|
|
|
|
* "bad retransmit" recovery
|
|
|
|
*/
|
|
|
|
if (tp->t_rxtshift == 1 &&
|
|
|
|
ticks < tp->t_badrxtwin) {
|
|
|
|
tp->snd_cwnd = tp->snd_cwnd_prev;
|
|
|
|
tp->snd_ssthresh =
|
|
|
|
tp->snd_ssthresh_prev;
|
|
|
|
tp->snd_nxt = tp->snd_max;
|
|
|
|
tp->t_badrxtwin = 0;
|
|
|
|
}
|
2001-11-22 04:50:44 +00:00
|
|
|
if ((to.to_flags & TOF_TS) != 0)
|
1995-02-09 23:13:27 +00:00
|
|
|
tcp_xmit_timer(tp,
|
1999-08-30 21:17:07 +00:00
|
|
|
ticks - to.to_tsecr + 1);
|
|
|
|
else if (tp->t_rtttime &&
|
2000-01-09 19:17:30 +00:00
|
|
|
SEQ_GT(th->th_ack, tp->t_rtseq))
|
1999-08-30 21:17:07 +00:00
|
|
|
tcp_xmit_timer(tp, ticks - tp->t_rtttime);
|
2000-01-09 19:17:30 +00:00
|
|
|
acked = th->th_ack - tp->snd_una;
|
1994-05-24 10:09:53 +00:00
|
|
|
tcpstat.tcps_rcvackpack++;
|
|
|
|
tcpstat.tcps_rcvackbyte += acked;
|
|
|
|
sbdrop(&so->so_snd, acked);
|
2000-01-09 19:17:30 +00:00
|
|
|
tp->snd_una = th->th_ack;
|
1994-05-24 10:09:53 +00:00
|
|
|
m_freem(m);
|
2000-01-09 19:17:30 +00:00
|
|
|
ND6_HINT(tp); /* some progress has been done */
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* 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)
|
1999-08-30 21:17:07 +00:00
|
|
|
callout_stop(tp->tt_rexmt);
|
|
|
|
else if (!callout_active(tp->tt_persist))
|
|
|
|
callout_reset(tp->tt_rexmt,
|
|
|
|
tp->t_rxtcur,
|
|
|
|
tcp_timer_rexmt, tp);
|
1994-05-24 10:09:53 +00:00
|
|
|
|
1998-05-31 18:42:49 +00:00
|
|
|
sowwakeup(so);
|
1994-05-24 10:09:53 +00:00
|
|
|
if (so->so_snd.sb_cc)
|
|
|
|
(void) tcp_output(tp);
|
|
|
|
return;
|
|
|
|
}
|
2000-01-09 19:17:30 +00:00
|
|
|
} else if (th->th_ack == tp->snd_una &&
|
|
|
|
LIST_EMPTY(&tp->t_segq) &&
|
|
|
|
tlen <= sbspace(&so->so_rcv)) {
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* 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;
|
2000-01-09 19:17:30 +00:00
|
|
|
tp->rcv_nxt += tlen;
|
1994-05-24 10:09:53 +00:00
|
|
|
tcpstat.tcps_rcvpack++;
|
2000-01-09 19:17:30 +00:00
|
|
|
tcpstat.tcps_rcvbyte += tlen;
|
|
|
|
ND6_HINT(tp); /* some progress has been done */
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
1995-04-05 10:32:14 +00:00
|
|
|
* Add data to socket buffer.
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
m_adj(m, drop_hdrlen); /* delayed header drop */
|
1994-05-24 10:09:53 +00:00
|
|
|
sbappend(&so->so_rcv, m);
|
|
|
|
sorwakeup(so);
|
2001-02-25 15:17:24 +00:00
|
|
|
if (DELAY_ACK(tp)) {
|
1999-08-30 21:17:07 +00:00
|
|
|
callout_reset(tp->tt_delack, tcp_delacktime,
|
|
|
|
tcp_timer_delack, tp);
|
1998-02-26 05:25:39 +00:00
|
|
|
} else {
|
1995-03-27 07:12:24 +00:00
|
|
|
tp->t_flags |= TF_ACKNOW;
|
|
|
|
tcp_output(tp);
|
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* 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;
|
1997-07-01 05:42:16 +00:00
|
|
|
tp->rcv_wnd = imax(win, (int)(tp->rcv_adv - tp->rcv_nxt));
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
switch (tp->t_state) {
|
|
|
|
|
1998-01-21 02:05:59 +00:00
|
|
|
/*
|
|
|
|
* If the state is SYN_RECEIVED:
|
|
|
|
* if seg contains an ACK, but not for our SYN/ACK, send a RST.
|
|
|
|
*/
|
|
|
|
case TCPS_SYN_RECEIVED:
|
2000-01-09 19:17:30 +00:00
|
|
|
if ((thflags & TH_ACK) &&
|
|
|
|
(SEQ_LEQ(th->th_ack, tp->snd_una) ||
|
2001-02-11 07:39:51 +00:00
|
|
|
SEQ_GT(th->th_ack, tp->snd_max))) {
|
|
|
|
rstreason = BANDLIM_RST_OPENPORT;
|
|
|
|
goto dropwithreset;
|
|
|
|
}
|
1998-01-21 02:05:59 +00:00
|
|
|
break;
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* 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:
|
2001-11-22 04:50:44 +00:00
|
|
|
if ((taop = tcp_gettaocache(&inp->inp_inc)) == NULL) {
|
1995-02-09 23:13:27 +00:00
|
|
|
taop = &tao_noncached;
|
|
|
|
bzero(taop, sizeof(*taop));
|
|
|
|
}
|
|
|
|
|
2000-01-09 19:17:30 +00:00
|
|
|
if ((thflags & TH_ACK) &&
|
|
|
|
(SEQ_LEQ(th->th_ack, tp->iss) ||
|
|
|
|
SEQ_GT(th->th_ack, tp->snd_max))) {
|
1995-02-09 23:13:27 +00:00
|
|
|
/*
|
|
|
|
* If we have a cached CCsent for the remote host,
|
|
|
|
* hence we haven't just crashed and restarted,
|
|
|
|
* do not send a RST. This may be a retransmission
|
|
|
|
* from the other side after our earlier ACK was lost.
|
|
|
|
* Our new SYN, when it arrives, will serve as the
|
|
|
|
* needed ACK.
|
|
|
|
*/
|
|
|
|
if (taop->tao_ccsent != 0)
|
|
|
|
goto drop;
|
2001-02-11 07:39:51 +00:00
|
|
|
else {
|
|
|
|
rstreason = BANDLIM_UNLIMITED;
|
1995-02-09 23:13:27 +00:00
|
|
|
goto dropwithreset;
|
2001-02-11 07:39:51 +00:00
|
|
|
}
|
1995-02-09 23:13:27 +00:00
|
|
|
}
|
2000-01-09 19:17:30 +00:00
|
|
|
if (thflags & TH_RST) {
|
|
|
|
if (thflags & TH_ACK)
|
1994-05-24 10:09:53 +00:00
|
|
|
tp = tcp_drop(tp, ECONNREFUSED);
|
|
|
|
goto drop;
|
|
|
|
}
|
2000-01-09 19:17:30 +00:00
|
|
|
if ((thflags & TH_SYN) == 0)
|
1994-05-24 10:09:53 +00:00
|
|
|
goto drop;
|
2000-01-09 19:17:30 +00:00
|
|
|
tp->snd_wnd = th->th_win; /* initial send window */
|
1995-02-09 23:13:27 +00:00
|
|
|
tp->cc_recv = to.to_cc; /* foreign CC */
|
|
|
|
|
2000-01-09 19:17:30 +00:00
|
|
|
tp->irs = th->th_seq;
|
1994-05-24 10:09:53 +00:00
|
|
|
tcp_rcvseqinit(tp);
|
2000-01-09 19:17:30 +00:00
|
|
|
if (thflags & TH_ACK) {
|
1995-02-09 23:13:27 +00:00
|
|
|
/*
|
|
|
|
* Our SYN was acked. If segment contains CC.ECHO
|
|
|
|
* option, check it to make sure this segment really
|
|
|
|
* matches our SYN. If not, just drop it as old
|
|
|
|
* duplicate, but send an RST if we're still playing
|
1997-09-30 16:38:09 +00:00
|
|
|
* by the old rules. If no CC.ECHO option, make sure
|
|
|
|
* we don't get fooled into using T/TCP.
|
1995-02-09 23:13:27 +00:00
|
|
|
*/
|
2001-11-22 04:50:44 +00:00
|
|
|
if (to.to_flags & TOF_CCECHO) {
|
1999-05-06 18:13:11 +00:00
|
|
|
if (tp->cc_send != to.to_ccecho) {
|
1997-09-30 16:38:09 +00:00
|
|
|
if (taop->tao_ccsent != 0)
|
|
|
|
goto drop;
|
2001-02-11 07:39:51 +00:00
|
|
|
else {
|
|
|
|
rstreason = BANDLIM_UNLIMITED;
|
1997-09-30 16:38:09 +00:00
|
|
|
goto dropwithreset;
|
2001-02-11 07:39:51 +00:00
|
|
|
}
|
1999-05-06 18:13:11 +00:00
|
|
|
}
|
1997-09-30 16:38:09 +00:00
|
|
|
} else
|
|
|
|
tp->t_flags &= ~TF_RCVD_CC;
|
1995-11-03 22:31:54 +00:00
|
|
|
tcpstat.tcps_connects++;
|
|
|
|
soisconnected(so);
|
|
|
|
/* 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;
|
|
|
|
}
|
1995-02-09 23:13:27 +00:00
|
|
|
/* Segment is acceptable, update cache if undefined. */
|
|
|
|
if (taop->tao_ccsent == 0)
|
|
|
|
taop->tao_ccsent = to.to_ccecho;
|
1995-05-30 08:16:23 +00:00
|
|
|
|
1995-02-09 23:13:27 +00:00
|
|
|
tp->rcv_adv += tp->rcv_wnd;
|
|
|
|
tp->snd_una++; /* SYN is acked */
|
|
|
|
/*
|
|
|
|
* If there's data, delay ACK; if there's also a FIN
|
|
|
|
* ACKNOW will be turned on later.
|
|
|
|
*/
|
2001-02-25 15:17:24 +00:00
|
|
|
if (DELAY_ACK(tp) && tlen != 0)
|
1999-08-30 21:17:07 +00:00
|
|
|
callout_reset(tp->tt_delack, tcp_delacktime,
|
|
|
|
tcp_timer_delack, tp);
|
1995-02-09 23:13:27 +00:00
|
|
|
else
|
|
|
|
tp->t_flags |= TF_ACKNOW;
|
|
|
|
/*
|
|
|
|
* Received <SYN,ACK> in SYN_SENT[*] state.
|
|
|
|
* Transitions:
|
|
|
|
* SYN_SENT --> ESTABLISHED
|
|
|
|
* SYN_SENT* --> FIN_WAIT_1
|
|
|
|
*/
|
1999-08-30 21:17:07 +00:00
|
|
|
tp->t_starttime = ticks;
|
1995-02-09 23:13:27 +00:00
|
|
|
if (tp->t_flags & TF_NEEDFIN) {
|
|
|
|
tp->t_state = TCPS_FIN_WAIT_1;
|
|
|
|
tp->t_flags &= ~TF_NEEDFIN;
|
2000-01-09 19:17:30 +00:00
|
|
|
thflags &= ~TH_SYN;
|
1996-09-13 18:47:03 +00:00
|
|
|
} else {
|
1995-02-09 23:13:27 +00:00
|
|
|
tp->t_state = TCPS_ESTABLISHED;
|
1999-08-30 21:17:07 +00:00
|
|
|
callout_reset(tp->tt_keep, tcp_keepidle,
|
|
|
|
tcp_timer_keep, tp);
|
1996-09-13 18:47:03 +00:00
|
|
|
}
|
1995-02-09 23:13:27 +00:00
|
|
|
} else {
|
|
|
|
/*
|
|
|
|
* Received initial SYN in SYN-SENT[*] state => simul-
|
|
|
|
* taneous open. If segment contains CC option and there is
|
1995-05-30 08:16:23 +00:00
|
|
|
* a cached CC, apply TAO test; if it succeeds, connection is
|
1995-02-09 23:13:27 +00:00
|
|
|
* half-synchronized. Otherwise, do 3-way handshake:
|
|
|
|
* SYN-SENT -> SYN-RECEIVED
|
|
|
|
* SYN-SENT* -> SYN-RECEIVED*
|
|
|
|
* If there was no CC option, clear cached CC value.
|
|
|
|
*/
|
|
|
|
tp->t_flags |= TF_ACKNOW;
|
1999-08-30 21:17:07 +00:00
|
|
|
callout_stop(tp->tt_rexmt);
|
2001-11-22 04:50:44 +00:00
|
|
|
if (to.to_flags & TOF_CC) {
|
1995-02-09 23:13:27 +00:00
|
|
|
if (taop->tao_cc != 0 &&
|
|
|
|
CC_GT(to.to_cc, taop->tao_cc)) {
|
|
|
|
/*
|
1995-05-30 08:16:23 +00:00
|
|
|
* update cache and make transition:
|
1995-02-09 23:13:27 +00:00
|
|
|
* SYN-SENT -> ESTABLISHED*
|
|
|
|
* SYN-SENT* -> FIN-WAIT-1*
|
|
|
|
*/
|
|
|
|
taop->tao_cc = to.to_cc;
|
1999-08-30 21:17:07 +00:00
|
|
|
tp->t_starttime = ticks;
|
1995-02-09 23:13:27 +00:00
|
|
|
if (tp->t_flags & TF_NEEDFIN) {
|
|
|
|
tp->t_state = TCPS_FIN_WAIT_1;
|
|
|
|
tp->t_flags &= ~TF_NEEDFIN;
|
1996-09-13 18:47:03 +00:00
|
|
|
} else {
|
1995-02-09 23:13:27 +00:00
|
|
|
tp->t_state = TCPS_ESTABLISHED;
|
1999-08-30 21:17:07 +00:00
|
|
|
callout_reset(tp->tt_keep,
|
|
|
|
tcp_keepidle,
|
|
|
|
tcp_timer_keep,
|
|
|
|
tp);
|
1996-09-13 18:47:03 +00:00
|
|
|
}
|
1995-02-09 23:13:27 +00:00
|
|
|
tp->t_flags |= TF_NEEDSYN;
|
|
|
|
} else
|
|
|
|
tp->t_state = TCPS_SYN_RECEIVED;
|
|
|
|
} else {
|
|
|
|
/* CC.NEW or no option => invalidate cache */
|
|
|
|
taop->tao_cc = 0;
|
|
|
|
tp->t_state = TCPS_SYN_RECEIVED;
|
|
|
|
}
|
1995-05-30 08:16:23 +00:00
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
trimthenstep6:
|
|
|
|
/*
|
2000-01-09 19:17:30 +00:00
|
|
|
* Advance th->th_seq to correspond to first data byte.
|
1994-05-24 10:09:53 +00:00
|
|
|
* If data, trim to stay within window,
|
|
|
|
* dropping FIN if necessary.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
th->th_seq++;
|
|
|
|
if (tlen > tp->rcv_wnd) {
|
|
|
|
todrop = tlen - tp->rcv_wnd;
|
1994-05-24 10:09:53 +00:00
|
|
|
m_adj(m, -todrop);
|
2000-01-09 19:17:30 +00:00
|
|
|
tlen = tp->rcv_wnd;
|
|
|
|
thflags &= ~TH_FIN;
|
1994-05-24 10:09:53 +00:00
|
|
|
tcpstat.tcps_rcvpackafterwin++;
|
|
|
|
tcpstat.tcps_rcvbyteafterwin += todrop;
|
|
|
|
}
|
2000-01-09 19:17:30 +00:00
|
|
|
tp->snd_wl1 = th->th_seq - 1;
|
|
|
|
tp->rcv_up = th->th_seq;
|
1995-02-09 23:13:27 +00:00
|
|
|
/*
|
|
|
|
* Client side of transaction: already sent SYN and data.
|
|
|
|
* If the remote host used T/TCP to validate the SYN,
|
|
|
|
* our data will be ACK'd; if so, enter normal data segment
|
|
|
|
* processing in the middle of step 5, ack processing.
|
|
|
|
* Otherwise, goto step 6.
|
1995-05-30 08:16:23 +00:00
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if (thflags & TH_ACK)
|
1995-02-09 23:13:27 +00:00
|
|
|
goto process_ACK;
|
1994-05-24 10:09:53 +00:00
|
|
|
goto step6;
|
1995-02-09 23:13:27 +00:00
|
|
|
/*
|
|
|
|
* If the state is LAST_ACK or CLOSING or TIME_WAIT:
|
|
|
|
* if segment contains a SYN and CC [not CC.NEW] option:
|
|
|
|
* if state == TIME_WAIT and connection duration > MSL,
|
|
|
|
* drop packet and send RST;
|
|
|
|
*
|
|
|
|
* if SEG.CC > CCrecv then is new SYN, and can implicitly
|
|
|
|
* ack the FIN (and data) in retransmission queue.
|
|
|
|
* Complete close and delete TCPCB. Then reprocess
|
|
|
|
* segment, hoping to find new TCPCB in LISTEN state;
|
|
|
|
*
|
|
|
|
* else must be old SYN; drop it.
|
1995-05-30 08:16:23 +00:00
|
|
|
* else do normal processing.
|
|
|
|
*/
|
1995-02-09 23:13:27 +00:00
|
|
|
case TCPS_LAST_ACK:
|
|
|
|
case TCPS_CLOSING:
|
|
|
|
case TCPS_TIME_WAIT:
|
2000-01-09 19:17:30 +00:00
|
|
|
if ((thflags & TH_SYN) &&
|
2001-11-22 04:50:44 +00:00
|
|
|
(to.to_flags & TOF_CC) && tp->cc_recv != 0) {
|
1995-02-09 23:13:27 +00:00
|
|
|
if (tp->t_state == TCPS_TIME_WAIT &&
|
2001-02-11 07:39:51 +00:00
|
|
|
(ticks - tp->t_starttime) > tcp_msl) {
|
|
|
|
rstreason = BANDLIM_UNLIMITED;
|
1995-02-09 23:13:27 +00:00
|
|
|
goto dropwithreset;
|
2001-02-11 07:39:51 +00:00
|
|
|
}
|
1995-02-09 23:13:27 +00:00
|
|
|
if (CC_GT(to.to_cc, tp->cc_recv)) {
|
|
|
|
tp = tcp_close(tp);
|
|
|
|
goto findpcb;
|
|
|
|
}
|
|
|
|
else
|
|
|
|
goto drop;
|
|
|
|
}
|
|
|
|
break; /* continue normal processing */
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* States other than LISTEN or SYN_SENT.
|
1998-09-11 16:04:03 +00:00
|
|
|
* First check the RST flag and sequence number since reset segments
|
|
|
|
* are exempt from the timestamp and connection count tests. This
|
|
|
|
* fixes a bug introduced by the Stevens, vol. 2, p. 960 bugfix
|
|
|
|
* below which allowed reset segments in half the sequence space
|
|
|
|
* to fall though and be processed (which gives forged reset
|
|
|
|
* segments with a random sequence number a 50 percent chance of
|
|
|
|
* killing a connection).
|
|
|
|
* Then check timestamp, if present.
|
1995-02-09 23:13:27 +00:00
|
|
|
* Then check the connection count, if present.
|
1995-05-30 08:16:23 +00:00
|
|
|
* Then check that at least some bytes of segment are within
|
1994-05-24 10:09:53 +00:00
|
|
|
* receive window. If segment begins before rcv_nxt,
|
|
|
|
* drop leading data (and SYN); if nothing left, just ack.
|
1995-02-09 23:13:27 +00:00
|
|
|
*
|
1998-09-11 16:04:03 +00:00
|
|
|
*
|
|
|
|
* If the RST bit is set, check the sequence number to see
|
|
|
|
* if this is a valid reset segment.
|
|
|
|
* RFC 793 page 37:
|
|
|
|
* In all states except SYN-SENT, all reset (RST) segments
|
|
|
|
* are validated by checking their SEQ-fields. A reset is
|
|
|
|
* valid if its sequence number is in the window.
|
|
|
|
* Note: this does not take into account delayed ACKs, so
|
|
|
|
* we should test against last_ack_sent instead of rcv_nxt.
|
1999-12-11 04:05:52 +00:00
|
|
|
* The sequence number in the reset segment is normally an
|
|
|
|
* echo of our outgoing acknowlegement numbers, but some hosts
|
|
|
|
* send a reset with the sequence number at the rightmost edge
|
|
|
|
* of our receive window, and we have to handle this case.
|
1998-09-11 16:04:03 +00:00
|
|
|
* If we have multiple segments in flight, the intial reset
|
|
|
|
* segment sequence numbers will be to the left of last_ack_sent,
|
|
|
|
* but they will eventually catch up.
|
|
|
|
* In any case, it never made sense to trim reset segments to
|
|
|
|
* fit the receive window since RFC 1122 says:
|
|
|
|
* 4.2.2.12 RST Segment: RFC-793 Section 3.4
|
|
|
|
*
|
|
|
|
* A TCP SHOULD allow a received RST segment to include data.
|
|
|
|
*
|
|
|
|
* DISCUSSION
|
|
|
|
* It has been suggested that a RST segment could contain
|
|
|
|
* ASCII text that encoded and explained the cause of the
|
|
|
|
* RST. No standard has yet been established for such
|
|
|
|
* data.
|
|
|
|
*
|
|
|
|
* If the reset segment passes the sequence number test examine
|
|
|
|
* the state:
|
|
|
|
* SYN_RECEIVED STATE:
|
|
|
|
* If passive open, return to LISTEN state.
|
|
|
|
* If active open, inform user that connection was refused.
|
2001-08-23 22:34:29 +00:00
|
|
|
* ESTABLISHED, FIN_WAIT_1, FIN_WAIT_2, CLOSE_WAIT STATES:
|
1998-09-11 16:04:03 +00:00
|
|
|
* Inform user that connection was reset, and close tcb.
|
1999-07-18 14:42:48 +00:00
|
|
|
* CLOSING, LAST_ACK STATES:
|
1998-09-11 16:04:03 +00:00
|
|
|
* Close the tcb.
|
1999-07-18 14:42:48 +00:00
|
|
|
* TIME_WAIT STATE:
|
1998-09-11 16:04:03 +00:00
|
|
|
* Drop the segment - see Stevens, vol. 2, p. 964 and
|
|
|
|
* RFC 1337.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if (thflags & TH_RST) {
|
|
|
|
if (SEQ_GEQ(th->th_seq, tp->last_ack_sent) &&
|
|
|
|
SEQ_LT(th->th_seq, tp->last_ack_sent + tp->rcv_wnd)) {
|
1998-09-11 16:04:03 +00:00
|
|
|
switch (tp->t_state) {
|
|
|
|
|
|
|
|
case TCPS_SYN_RECEIVED:
|
|
|
|
so->so_error = ECONNREFUSED;
|
|
|
|
goto close;
|
|
|
|
|
|
|
|
case TCPS_ESTABLISHED:
|
|
|
|
case TCPS_FIN_WAIT_1:
|
|
|
|
case TCPS_FIN_WAIT_2:
|
|
|
|
case TCPS_CLOSE_WAIT:
|
|
|
|
so->so_error = ECONNRESET;
|
|
|
|
close:
|
|
|
|
tp->t_state = TCPS_CLOSED;
|
|
|
|
tcpstat.tcps_drops++;
|
|
|
|
tp = tcp_close(tp);
|
|
|
|
break;
|
|
|
|
|
|
|
|
case TCPS_CLOSING:
|
|
|
|
case TCPS_LAST_ACK:
|
|
|
|
tp = tcp_close(tp);
|
|
|
|
break;
|
|
|
|
|
|
|
|
case TCPS_TIME_WAIT:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
goto drop;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
1994-05-24 10:09:53 +00:00
|
|
|
* RFC 1323 PAWS: If we have a timestamp reply on this segment
|
|
|
|
* and it's less than ts_recent, drop it.
|
|
|
|
*/
|
2001-11-22 04:50:44 +00:00
|
|
|
if ((to.to_flags & TOF_TS) != 0 && tp->ts_recent &&
|
1998-09-11 16:04:03 +00:00
|
|
|
TSTMP_LT(to.to_tsval, tp->ts_recent)) {
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
/* Check to see if ts_recent is over 24 days old. */
|
1999-08-30 21:17:07 +00:00
|
|
|
if ((int)(ticks - tp->ts_recent_age) > TCP_PAWS_IDLE) {
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* 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++;
|
2000-01-09 19:17:30 +00:00
|
|
|
tcpstat.tcps_rcvdupbyte += tlen;
|
1994-05-24 10:09:53 +00:00
|
|
|
tcpstat.tcps_pawsdrop++;
|
|
|
|
goto dropafterack;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
1995-02-09 23:13:27 +00:00
|
|
|
/*
|
|
|
|
* T/TCP mechanism
|
|
|
|
* If T/TCP was negotiated and the segment doesn't have CC,
|
1998-04-17 22:37:19 +00:00
|
|
|
* or if its CC is wrong then drop the segment.
|
1995-02-09 23:13:27 +00:00
|
|
|
* RST segments do not have to comply with this.
|
|
|
|
*/
|
|
|
|
if ((tp->t_flags & (TF_REQ_CC|TF_RCVD_CC)) == (TF_REQ_CC|TF_RCVD_CC) &&
|
2001-11-22 04:50:44 +00:00
|
|
|
((to.to_flags & TOF_CC) == 0 || tp->cc_recv != to.to_cc))
|
1995-02-09 23:13:27 +00:00
|
|
|
goto dropafterack;
|
|
|
|
|
1998-09-11 16:04:03 +00:00
|
|
|
/*
|
|
|
|
* In the SYN-RECEIVED state, validate that the packet belongs to
|
|
|
|
* this connection before trimming the data to fit the receive
|
|
|
|
* window. Check the sequence number versus IRS since we know
|
|
|
|
* the sequence numbers haven't wrapped. This is a partial fix
|
|
|
|
* for the "LAND" DoS attack.
|
|
|
|
*/
|
2001-02-11 07:39:51 +00:00
|
|
|
if (tp->t_state == TCPS_SYN_RECEIVED && SEQ_LT(th->th_seq, tp->irs)) {
|
|
|
|
rstreason = BANDLIM_RST_OPENPORT;
|
|
|
|
goto dropwithreset;
|
|
|
|
}
|
1998-09-11 16:04:03 +00:00
|
|
|
|
2000-01-09 19:17:30 +00:00
|
|
|
todrop = tp->rcv_nxt - th->th_seq;
|
1994-05-24 10:09:53 +00:00
|
|
|
if (todrop > 0) {
|
2000-01-09 19:17:30 +00:00
|
|
|
if (thflags & TH_SYN) {
|
|
|
|
thflags &= ~TH_SYN;
|
|
|
|
th->th_seq++;
|
|
|
|
if (th->th_urp > 1)
|
|
|
|
th->th_urp--;
|
1994-05-24 10:09:53 +00:00
|
|
|
else
|
2000-01-09 19:17:30 +00:00
|
|
|
thflags &= ~TH_URG;
|
1994-05-24 10:09:53 +00:00
|
|
|
todrop--;
|
|
|
|
}
|
1995-02-16 01:39:19 +00:00
|
|
|
/*
|
|
|
|
* Following if statement from Stevens, vol. 2, p. 960.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if (todrop > tlen
|
|
|
|
|| (todrop == tlen && (thflags & TH_FIN) == 0)) {
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
1995-02-16 01:39:19 +00:00
|
|
|
* 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.
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
thflags &= ~TH_FIN;
|
1995-02-16 01:39:19 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Send an ACK to resynchronize and drop any data.
|
|
|
|
* But keep on processing for RST or ACK.
|
|
|
|
*/
|
|
|
|
tp->t_flags |= TF_ACKNOW;
|
2000-01-09 19:17:30 +00:00
|
|
|
todrop = tlen;
|
1995-02-16 01:39:19 +00:00
|
|
|
tcpstat.tcps_rcvduppack++;
|
|
|
|
tcpstat.tcps_rcvdupbyte += todrop;
|
1994-05-24 10:09:53 +00:00
|
|
|
} else {
|
|
|
|
tcpstat.tcps_rcvpartduppack++;
|
|
|
|
tcpstat.tcps_rcvpartdupbyte += todrop;
|
|
|
|
}
|
2000-01-09 19:17:30 +00:00
|
|
|
drop_hdrlen += todrop; /* drop from the top afterwards */
|
|
|
|
th->th_seq += todrop;
|
|
|
|
tlen -= todrop;
|
|
|
|
if (th->th_urp > todrop)
|
|
|
|
th->th_urp -= todrop;
|
1994-05-24 10:09:53 +00:00
|
|
|
else {
|
2000-01-09 19:17:30 +00:00
|
|
|
thflags &= ~TH_URG;
|
|
|
|
th->th_urp = 0;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* 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) &&
|
2000-01-09 19:17:30 +00:00
|
|
|
tp->t_state > TCPS_CLOSE_WAIT && tlen) {
|
1994-05-24 10:09:53 +00:00
|
|
|
tp = tcp_close(tp);
|
|
|
|
tcpstat.tcps_rcvafterclose++;
|
2001-02-11 07:39:51 +00:00
|
|
|
rstreason = BANDLIM_UNLIMITED;
|
1994-05-24 10:09:53 +00:00
|
|
|
goto dropwithreset;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If segment ends after window, drop trailing data
|
|
|
|
* (and PUSH and FIN); if nothing left, just ACK.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
todrop = (th->th_seq+tlen) - (tp->rcv_nxt+tp->rcv_wnd);
|
1994-05-24 10:09:53 +00:00
|
|
|
if (todrop > 0) {
|
|
|
|
tcpstat.tcps_rcvpackafterwin++;
|
2000-01-09 19:17:30 +00:00
|
|
|
if (todrop >= tlen) {
|
|
|
|
tcpstat.tcps_rcvbyteafterwin += tlen;
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* If a new connection request is received
|
|
|
|
* while in TIME_WAIT, drop the old connection
|
|
|
|
* and start over if the sequence numbers
|
|
|
|
* are above the previous ones.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if (thflags & TH_SYN &&
|
1994-05-24 10:09:53 +00:00
|
|
|
tp->t_state == TCPS_TIME_WAIT &&
|
2000-01-09 19:17:30 +00:00
|
|
|
SEQ_GT(th->th_seq, tp->rcv_nxt)) {
|
2001-08-22 00:58:16 +00:00
|
|
|
iss = tcp_new_isn(tp);
|
1994-05-24 10:09:53 +00:00
|
|
|
tp = tcp_close(tp);
|
|
|
|
goto findpcb;
|
|
|
|
}
|
|
|
|
/*
|
|
|
|
* If window is closed can only take segments at
|
|
|
|
* window edge, and have to drop data and PUSH from
|
|
|
|
* incoming segments. Continue processing, but
|
|
|
|
* remember to ack. Otherwise, drop segment
|
|
|
|
* and ack.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if (tp->rcv_wnd == 0 && th->th_seq == tp->rcv_nxt) {
|
1994-05-24 10:09:53 +00:00
|
|
|
tp->t_flags |= TF_ACKNOW;
|
|
|
|
tcpstat.tcps_rcvwinprobe++;
|
|
|
|
} else
|
|
|
|
goto dropafterack;
|
|
|
|
} else
|
|
|
|
tcpstat.tcps_rcvbyteafterwin += todrop;
|
|
|
|
m_adj(m, -todrop);
|
2000-01-09 19:17:30 +00:00
|
|
|
tlen -= todrop;
|
|
|
|
thflags &= ~(TH_PUSH|TH_FIN);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If last ACK falls within this segment's sequence numbers,
|
|
|
|
* record its timestamp.
|
1995-02-09 23:13:27 +00:00
|
|
|
* NOTE that the test is modified according to the latest
|
|
|
|
* proposal of the tcplw@cray.com list (Braden 1993/04/26).
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
2001-11-22 04:50:44 +00:00
|
|
|
if ((to.to_flags & TOF_TS) != 0 &&
|
2000-01-09 19:17:30 +00:00
|
|
|
SEQ_LEQ(th->th_seq, tp->last_ack_sent)) {
|
1999-08-30 21:17:07 +00:00
|
|
|
tp->ts_recent_age = ticks;
|
1995-02-09 23:13:27 +00:00
|
|
|
tp->ts_recent = to.to_tsval;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If a SYN is in the window, then this is an
|
|
|
|
* error and we send an RST and drop the connection.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if (thflags & TH_SYN) {
|
1994-05-24 10:09:53 +00:00
|
|
|
tp = tcp_drop(tp, ECONNRESET);
|
2001-02-11 07:39:51 +00:00
|
|
|
rstreason = BANDLIM_UNLIMITED;
|
1994-05-24 10:09:53 +00:00
|
|
|
goto dropwithreset;
|
|
|
|
}
|
|
|
|
|
1995-02-09 23:13:27 +00:00
|
|
|
/*
|
|
|
|
* If the ACK bit is off: if in SYN-RECEIVED state or SENDSYN
|
|
|
|
* flag is on (half-synchronized state), then queue data for
|
|
|
|
* later processing; else drop segment and return.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if ((thflags & TH_ACK) == 0) {
|
1995-02-09 23:13:27 +00:00
|
|
|
if (tp->t_state == TCPS_SYN_RECEIVED ||
|
|
|
|
(tp->t_flags & TF_NEEDSYN))
|
|
|
|
goto step6;
|
|
|
|
else
|
|
|
|
goto drop;
|
|
|
|
}
|
1995-05-30 08:16:23 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Ack processing.
|
|
|
|
*/
|
|
|
|
switch (tp->t_state) {
|
|
|
|
|
|
|
|
/*
|
1998-01-21 02:05:59 +00:00
|
|
|
* In SYN_RECEIVED state, the ack ACKs our SYN, so enter
|
|
|
|
* ESTABLISHED state and continue processing.
|
|
|
|
* The ACK was checked above.
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
|
|
|
case TCPS_SYN_RECEIVED:
|
1995-02-09 23:13:27 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
tcpstat.tcps_connects++;
|
|
|
|
soisconnected(so);
|
|
|
|
/* Do window scaling? */
|
|
|
|
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;
|
|
|
|
}
|
1995-02-09 23:13:27 +00:00
|
|
|
/*
|
|
|
|
* Upon successful completion of 3-way handshake,
|
|
|
|
* update cache.CC if it was undefined, pass any queued
|
|
|
|
* data to the user, and advance state appropriately.
|
|
|
|
*/
|
2001-11-22 04:50:44 +00:00
|
|
|
if ((taop = tcp_gettaocache(&inp->inp_inc)) != NULL &&
|
1995-02-09 23:13:27 +00:00
|
|
|
taop->tao_cc == 0)
|
|
|
|
taop->tao_cc = tp->cc_recv;
|
|
|
|
|
|
|
|
/*
|
1995-05-30 08:16:23 +00:00
|
|
|
* Make transitions:
|
1995-02-09 23:13:27 +00:00
|
|
|
* SYN-RECEIVED -> ESTABLISHED
|
|
|
|
* SYN-RECEIVED* -> FIN-WAIT-1
|
|
|
|
*/
|
1999-08-30 21:17:07 +00:00
|
|
|
tp->t_starttime = ticks;
|
1995-02-09 23:13:27 +00:00
|
|
|
if (tp->t_flags & TF_NEEDFIN) {
|
|
|
|
tp->t_state = TCPS_FIN_WAIT_1;
|
|
|
|
tp->t_flags &= ~TF_NEEDFIN;
|
1996-09-13 18:47:03 +00:00
|
|
|
} else {
|
1995-02-09 23:13:27 +00:00
|
|
|
tp->t_state = TCPS_ESTABLISHED;
|
1999-08-30 21:17:07 +00:00
|
|
|
callout_reset(tp->tt_keep, tcp_keepidle,
|
|
|
|
tcp_timer_keep, tp);
|
1996-09-13 18:47:03 +00:00
|
|
|
}
|
1995-05-30 08:16:23 +00:00
|
|
|
/*
|
1995-02-09 23:13:27 +00:00
|
|
|
* If segment contains data or ACK, will call tcp_reass()
|
|
|
|
* later; if not, do so now to pass queued data to user.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if (tlen == 0 && (thflags & TH_FIN) == 0)
|
|
|
|
(void) tcp_reass(tp, (struct tcphdr *)0, 0,
|
1995-02-09 23:13:27 +00:00
|
|
|
(struct mbuf *)0);
|
2000-01-09 19:17:30 +00:00
|
|
|
tp->snd_wl1 = th->th_seq - 1;
|
1994-05-24 10:09:53 +00:00
|
|
|
/* fall into ... */
|
|
|
|
|
|
|
|
/*
|
|
|
|
* In ESTABLISHED state: drop duplicate ACKs; ACK out of range
|
|
|
|
* ACKs. If the ack is in the range
|
2000-01-09 19:17:30 +00:00
|
|
|
* tp->snd_una < th->th_ack <= tp->snd_max
|
|
|
|
* then advance tp->snd_una to th->th_ack and drop
|
1994-05-24 10:09:53 +00:00
|
|
|
* data from the retransmission queue. If this ACK reflects
|
|
|
|
* more up to date window information we update our window information.
|
|
|
|
*/
|
|
|
|
case TCPS_ESTABLISHED:
|
|
|
|
case TCPS_FIN_WAIT_1:
|
|
|
|
case TCPS_FIN_WAIT_2:
|
|
|
|
case TCPS_CLOSE_WAIT:
|
|
|
|
case TCPS_CLOSING:
|
|
|
|
case TCPS_LAST_ACK:
|
|
|
|
case TCPS_TIME_WAIT:
|
|
|
|
|
2000-01-09 19:17:30 +00:00
|
|
|
if (SEQ_LEQ(th->th_ack, tp->snd_una)) {
|
|
|
|
if (tlen == 0 && tiwin == tp->snd_wnd) {
|
1994-05-24 10:09:53 +00:00
|
|
|
tcpstat.tcps_rcvdupack++;
|
|
|
|
/*
|
|
|
|
* If we have outstanding data (other than
|
|
|
|
* a window probe), this is a completely
|
|
|
|
* duplicate ack (ie, window info didn't
|
|
|
|
* change), the ack is the biggest we've
|
|
|
|
* seen and we've seen exactly our rexmt
|
|
|
|
* threshhold of them, assume a packet
|
|
|
|
* has been dropped and retransmit it.
|
|
|
|
* Kludge snd_nxt & the congestion
|
|
|
|
* window so we send only this one
|
|
|
|
* packet.
|
|
|
|
*
|
|
|
|
* We know we're losing at the current
|
|
|
|
* window size so do congestion avoidance
|
|
|
|
* (set ssthresh to half the current window
|
|
|
|
* and pull our congestion window back to
|
|
|
|
* the new ssthresh).
|
|
|
|
*
|
|
|
|
* Dup acks mean that packets have left the
|
1995-05-30 08:16:23 +00:00
|
|
|
* network (they're now cached at the receiver)
|
1994-05-24 10:09:53 +00:00
|
|
|
* so bump cwnd by the amount in the receiver
|
|
|
|
* to keep a constant cwnd packets in the
|
|
|
|
* network.
|
|
|
|
*/
|
1999-08-30 21:17:07 +00:00
|
|
|
if (!callout_active(tp->tt_rexmt) ||
|
2000-01-09 19:17:30 +00:00
|
|
|
th->th_ack != tp->snd_una)
|
1994-05-24 10:09:53 +00:00
|
|
|
tp->t_dupacks = 0;
|
|
|
|
else if (++tp->t_dupacks == tcprexmtthresh) {
|
|
|
|
tcp_seq onxt = tp->snd_nxt;
|
|
|
|
u_int win =
|
|
|
|
min(tp->snd_wnd, tp->snd_cwnd) / 2 /
|
|
|
|
tp->t_maxseg;
|
2000-05-06 03:31:09 +00:00
|
|
|
if (tcp_do_newreno && SEQ_LT(th->th_ack,
|
|
|
|
tp->snd_recover)) {
|
|
|
|
/* False retransmit, should not
|
|
|
|
* cut window
|
|
|
|
*/
|
|
|
|
tp->snd_cwnd += tp->t_maxseg;
|
|
|
|
tp->t_dupacks = 0;
|
|
|
|
(void) tcp_output(tp);
|
|
|
|
goto drop;
|
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
if (win < 2)
|
|
|
|
win = 2;
|
|
|
|
tp->snd_ssthresh = win * tp->t_maxseg;
|
2000-05-06 03:31:09 +00:00
|
|
|
tp->snd_recover = tp->snd_max;
|
1999-08-30 21:17:07 +00:00
|
|
|
callout_stop(tp->tt_rexmt);
|
|
|
|
tp->t_rtttime = 0;
|
2000-01-09 19:17:30 +00:00
|
|
|
tp->snd_nxt = th->th_ack;
|
1994-05-24 10:09:53 +00:00
|
|
|
tp->snd_cwnd = tp->t_maxseg;
|
|
|
|
(void) tcp_output(tp);
|
|
|
|
tp->snd_cwnd = tp->snd_ssthresh +
|
|
|
|
tp->t_maxseg * tp->t_dupacks;
|
|
|
|
if (SEQ_GT(onxt, tp->snd_nxt))
|
|
|
|
tp->snd_nxt = onxt;
|
|
|
|
goto drop;
|
|
|
|
} else if (tp->t_dupacks > tcprexmtthresh) {
|
|
|
|
tp->snd_cwnd += tp->t_maxseg;
|
|
|
|
(void) tcp_output(tp);
|
|
|
|
goto drop;
|
|
|
|
}
|
|
|
|
} else
|
|
|
|
tp->t_dupacks = 0;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
/*
|
|
|
|
* If the congestion window was inflated to account
|
|
|
|
* for the other side's cached packets, retract it.
|
|
|
|
*/
|
2000-05-06 03:31:09 +00:00
|
|
|
if (tcp_do_newreno == 0) {
|
|
|
|
if (tp->t_dupacks >= tcprexmtthresh &&
|
|
|
|
tp->snd_cwnd > tp->snd_ssthresh)
|
|
|
|
tp->snd_cwnd = tp->snd_ssthresh;
|
|
|
|
tp->t_dupacks = 0;
|
|
|
|
} else if (tp->t_dupacks >= tcprexmtthresh &&
|
|
|
|
!tcp_newreno(tp, th)) {
|
|
|
|
/*
|
|
|
|
* Window inflation should have left us with approx.
|
|
|
|
* snd_ssthresh outstanding data. But in case we
|
|
|
|
* would be inclined to send a burst, better to do
|
|
|
|
* it via the slow start mechanism.
|
|
|
|
*/
|
|
|
|
if (SEQ_GT(th->th_ack + tp->snd_ssthresh, tp->snd_max))
|
|
|
|
tp->snd_cwnd =
|
|
|
|
tp->snd_max - th->th_ack + tp->t_maxseg;
|
|
|
|
else
|
|
|
|
tp->snd_cwnd = tp->snd_ssthresh;
|
|
|
|
tp->t_dupacks = 0;
|
|
|
|
}
|
2001-08-29 23:54:13 +00:00
|
|
|
if (tp->t_dupacks < tcprexmtthresh)
|
|
|
|
tp->t_dupacks = 0;
|
2000-01-09 19:17:30 +00:00
|
|
|
if (SEQ_GT(th->th_ack, tp->snd_max)) {
|
1994-05-24 10:09:53 +00:00
|
|
|
tcpstat.tcps_rcvacktoomuch++;
|
|
|
|
goto dropafterack;
|
|
|
|
}
|
1995-02-09 23:13:27 +00:00
|
|
|
/*
|
1995-05-30 08:16:23 +00:00
|
|
|
* If we reach this point, ACK is not a duplicate,
|
1995-02-09 23:13:27 +00:00
|
|
|
* i.e., it ACKs something we sent.
|
|
|
|
*/
|
|
|
|
if (tp->t_flags & TF_NEEDSYN) {
|
1995-05-30 08:16:23 +00:00
|
|
|
/*
|
1996-01-31 08:22:24 +00:00
|
|
|
* T/TCP: Connection was half-synchronized, and our
|
|
|
|
* SYN has been ACK'd (so connection is now fully
|
|
|
|
* synchronized). Go to non-starred state,
|
|
|
|
* increment snd_una for ACK of SYN, and check if
|
|
|
|
* we can do window scaling.
|
1995-02-09 23:13:27 +00:00
|
|
|
*/
|
|
|
|
tp->t_flags &= ~TF_NEEDSYN;
|
|
|
|
tp->snd_una++;
|
1996-01-31 08:22:24 +00:00
|
|
|
/* Do window scaling? */
|
|
|
|
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;
|
|
|
|
}
|
1995-02-09 23:13:27 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
process_ACK:
|
2000-01-09 19:17:30 +00:00
|
|
|
acked = th->th_ack - tp->snd_una;
|
1994-05-24 10:09:53 +00:00
|
|
|
tcpstat.tcps_rcvackpack++;
|
|
|
|
tcpstat.tcps_rcvackbyte += acked;
|
|
|
|
|
1999-08-30 21:17:07 +00:00
|
|
|
/*
|
|
|
|
* If we just performed our first retransmit, and the ACK
|
|
|
|
* arrives within our recovery window, then it was a mistake
|
|
|
|
* to do the retransmit in the first place. Recover our
|
|
|
|
* original cwnd and ssthresh, and proceed to transmit where
|
|
|
|
* we left off.
|
|
|
|
*/
|
|
|
|
if (tp->t_rxtshift == 1 && ticks < tp->t_badrxtwin) {
|
|
|
|
tp->snd_cwnd = tp->snd_cwnd_prev;
|
|
|
|
tp->snd_ssthresh = tp->snd_ssthresh_prev;
|
|
|
|
tp->snd_nxt = tp->snd_max;
|
|
|
|
tp->t_badrxtwin = 0; /* XXX probably not required */
|
|
|
|
}
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* If we have a timestamp reply, update smoothed
|
|
|
|
* round trip time. If no timestamp is present but
|
|
|
|
* transmit timer is running and timed sequence
|
|
|
|
* number was acked, update smoothed round trip time.
|
|
|
|
* Since we now have an rtt measurement, cancel the
|
|
|
|
* timer backoff (cf., Phil Karn's retransmit alg.).
|
|
|
|
* Recompute the initial retransmit timer.
|
|
|
|
*/
|
2001-11-22 04:50:44 +00:00
|
|
|
if (to.to_flags & TOF_TS)
|
1999-08-30 21:17:07 +00:00
|
|
|
tcp_xmit_timer(tp, ticks - to.to_tsecr + 1);
|
2000-01-09 19:17:30 +00:00
|
|
|
else if (tp->t_rtttime && SEQ_GT(th->th_ack, tp->t_rtseq))
|
1999-08-30 21:17:07 +00:00
|
|
|
tcp_xmit_timer(tp, ticks - tp->t_rtttime);
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* If all outstanding data is acked, stop retransmit
|
|
|
|
* timer and remember to restart (more output or persist).
|
|
|
|
* If there is more data to be acked, restart retransmit
|
|
|
|
* timer, using current (possibly backed-off) value.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if (th->th_ack == tp->snd_max) {
|
1999-08-30 21:17:07 +00:00
|
|
|
callout_stop(tp->tt_rexmt);
|
1994-05-24 10:09:53 +00:00
|
|
|
needoutput = 1;
|
1999-08-30 21:17:07 +00:00
|
|
|
} else if (!callout_active(tp->tt_persist))
|
|
|
|
callout_reset(tp->tt_rexmt, tp->t_rxtcur,
|
|
|
|
tcp_timer_rexmt, tp);
|
1995-02-09 23:13:27 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* If no data (only SYN) was ACK'd,
|
|
|
|
* skip rest of ACK processing.
|
|
|
|
*/
|
|
|
|
if (acked == 0)
|
|
|
|
goto step6;
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* When new data is acked, open the congestion window.
|
|
|
|
* If the window gives us less than ssthresh packets
|
|
|
|
* in flight, open exponentially (maxseg per packet).
|
|
|
|
* Otherwise open linearly: maxseg per window
|
1994-10-13 18:36:32 +00:00
|
|
|
* (maxseg^2 / cwnd per packet).
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
|
|
|
{
|
|
|
|
register u_int cw = tp->snd_cwnd;
|
|
|
|
register u_int incr = tp->t_maxseg;
|
|
|
|
|
|
|
|
if (cw > tp->snd_ssthresh)
|
1994-10-13 18:36:32 +00:00
|
|
|
incr = incr * incr / cw;
|
2000-11-04 15:59:39 +00:00
|
|
|
/*
|
|
|
|
* If t_dupacks != 0 here, it indicates that we are still
|
|
|
|
* in NewReno fast recovery mode, so we leave the congestion
|
|
|
|
* window alone.
|
|
|
|
*/
|
|
|
|
if (tcp_do_newreno == 0 || tp->t_dupacks == 0)
|
2000-05-06 03:31:09 +00:00
|
|
|
tp->snd_cwnd = min(cw + incr,TCP_MAXWIN<<tp->snd_scale);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
if (acked > so->so_snd.sb_cc) {
|
|
|
|
tp->snd_wnd -= so->so_snd.sb_cc;
|
|
|
|
sbdrop(&so->so_snd, (int)so->so_snd.sb_cc);
|
|
|
|
ourfinisacked = 1;
|
|
|
|
} else {
|
|
|
|
sbdrop(&so->so_snd, acked);
|
|
|
|
tp->snd_wnd -= acked;
|
|
|
|
ourfinisacked = 0;
|
|
|
|
}
|
1998-05-31 18:42:49 +00:00
|
|
|
sowwakeup(so);
|
2000-01-09 19:17:30 +00:00
|
|
|
tp->snd_una = th->th_ack;
|
1994-05-24 10:09:53 +00:00
|
|
|
if (SEQ_LT(tp->snd_nxt, tp->snd_una))
|
|
|
|
tp->snd_nxt = tp->snd_una;
|
|
|
|
|
|
|
|
switch (tp->t_state) {
|
|
|
|
|
|
|
|
/*
|
|
|
|
* In FIN_WAIT_1 STATE in addition to the processing
|
|
|
|
* for the ESTABLISHED state if our FIN is now acknowledged
|
|
|
|
* then enter FIN_WAIT_2.
|
|
|
|
*/
|
|
|
|
case TCPS_FIN_WAIT_1:
|
|
|
|
if (ourfinisacked) {
|
|
|
|
/*
|
|
|
|
* If we can't receive any more
|
|
|
|
* data, then closing user can proceed.
|
|
|
|
* Starting the timer is contrary to the
|
|
|
|
* specification, but if we don't get a FIN
|
|
|
|
* we'll hang forever.
|
|
|
|
*/
|
|
|
|
if (so->so_state & SS_CANTRCVMORE) {
|
|
|
|
soisdisconnected(so);
|
1999-08-30 21:17:07 +00:00
|
|
|
callout_reset(tp->tt_2msl, tcp_maxidle,
|
|
|
|
tcp_timer_2msl, tp);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
tp->t_state = TCPS_FIN_WAIT_2;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* In CLOSING STATE in addition to the processing for
|
|
|
|
* the ESTABLISHED state if the ACK acknowledges our FIN
|
|
|
|
* then enter the TIME-WAIT state, otherwise ignore
|
|
|
|
* the segment.
|
|
|
|
*/
|
|
|
|
case TCPS_CLOSING:
|
|
|
|
if (ourfinisacked) {
|
|
|
|
tp->t_state = TCPS_TIME_WAIT;
|
|
|
|
tcp_canceltimers(tp);
|
1995-02-09 23:13:27 +00:00
|
|
|
/* Shorten TIME_WAIT [RFC-1644, p.28] */
|
|
|
|
if (tp->cc_recv != 0 &&
|
1999-08-30 21:17:07 +00:00
|
|
|
(ticks - tp->t_starttime) < tcp_msl)
|
|
|
|
callout_reset(tp->tt_2msl,
|
|
|
|
tp->t_rxtcur *
|
|
|
|
TCPTV_TWTRUNC,
|
|
|
|
tcp_timer_2msl, tp);
|
1995-02-09 23:13:27 +00:00
|
|
|
else
|
1999-08-30 21:17:07 +00:00
|
|
|
callout_reset(tp->tt_2msl, 2 * tcp_msl,
|
|
|
|
tcp_timer_2msl, tp);
|
1994-05-24 10:09:53 +00:00
|
|
|
soisdisconnected(so);
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* In LAST_ACK, we may still be waiting for data to drain
|
|
|
|
* and/or to be acked, as well as for the ack of our FIN.
|
|
|
|
* If our FIN is now acknowledged, delete the TCB,
|
|
|
|
* enter the closed state and return.
|
|
|
|
*/
|
|
|
|
case TCPS_LAST_ACK:
|
|
|
|
if (ourfinisacked) {
|
|
|
|
tp = tcp_close(tp);
|
|
|
|
goto drop;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* In TIME_WAIT state the only thing that should arrive
|
|
|
|
* is a retransmission of the remote FIN. Acknowledge
|
|
|
|
* it and restart the finack timer.
|
|
|
|
*/
|
|
|
|
case TCPS_TIME_WAIT:
|
1999-08-30 21:17:07 +00:00
|
|
|
callout_reset(tp->tt_2msl, 2 * tcp_msl,
|
|
|
|
tcp_timer_2msl, tp);
|
1994-05-24 10:09:53 +00:00
|
|
|
goto dropafterack;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
step6:
|
|
|
|
/*
|
|
|
|
* Update window information.
|
|
|
|
* Don't look at window if no ACK: TAC's send garbage on first SYN.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if ((thflags & TH_ACK) &&
|
|
|
|
(SEQ_LT(tp->snd_wl1, th->th_seq) ||
|
|
|
|
(tp->snd_wl1 == th->th_seq && (SEQ_LT(tp->snd_wl2, th->th_ack) ||
|
|
|
|
(tp->snd_wl2 == th->th_ack && tiwin > tp->snd_wnd))))) {
|
1994-05-24 10:09:53 +00:00
|
|
|
/* keep track of pure window updates */
|
2000-01-09 19:17:30 +00:00
|
|
|
if (tlen == 0 &&
|
|
|
|
tp->snd_wl2 == th->th_ack && tiwin > tp->snd_wnd)
|
1994-05-24 10:09:53 +00:00
|
|
|
tcpstat.tcps_rcvwinupd++;
|
|
|
|
tp->snd_wnd = tiwin;
|
2000-01-09 19:17:30 +00:00
|
|
|
tp->snd_wl1 = th->th_seq;
|
|
|
|
tp->snd_wl2 = th->th_ack;
|
1994-05-24 10:09:53 +00:00
|
|
|
if (tp->snd_wnd > tp->max_sndwnd)
|
|
|
|
tp->max_sndwnd = tp->snd_wnd;
|
|
|
|
needoutput = 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Process segments with URG.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if ((thflags & TH_URG) && th->th_urp &&
|
1994-05-24 10:09:53 +00:00
|
|
|
TCPS_HAVERCVDFIN(tp->t_state) == 0) {
|
|
|
|
/*
|
|
|
|
* This is a kludge, but if we receive and accept
|
|
|
|
* random urgent pointers, we'll crash in
|
|
|
|
* soreceive. It's hard to imagine someone
|
|
|
|
* actually wanting to send this much urgent data.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if (th->th_urp + so->so_rcv.sb_cc > sb_max) {
|
|
|
|
th->th_urp = 0; /* XXX */
|
|
|
|
thflags &= ~TH_URG; /* XXX */
|
1994-05-24 10:09:53 +00:00
|
|
|
goto dodata; /* XXX */
|
|
|
|
}
|
|
|
|
/*
|
|
|
|
* If this segment advances the known urgent pointer,
|
|
|
|
* then mark the data stream. This should not happen
|
|
|
|
* in CLOSE_WAIT, CLOSING, LAST_ACK or TIME_WAIT STATES since
|
1995-05-30 08:16:23 +00:00
|
|
|
* a FIN has been received from the remote side.
|
1994-05-24 10:09:53 +00:00
|
|
|
* In these states we ignore the URG.
|
|
|
|
*
|
|
|
|
* According to RFC961 (Assigned Protocols),
|
|
|
|
* the urgent pointer points to the last octet
|
|
|
|
* of urgent data. We continue, however,
|
|
|
|
* to consider it to indicate the first octet
|
1995-05-30 08:16:23 +00:00
|
|
|
* of data past the urgent section as the original
|
1994-05-24 10:09:53 +00:00
|
|
|
* spec states (in one of two places).
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if (SEQ_GT(th->th_seq+th->th_urp, tp->rcv_up)) {
|
|
|
|
tp->rcv_up = th->th_seq + th->th_urp;
|
1994-05-24 10:09:53 +00:00
|
|
|
so->so_oobmark = so->so_rcv.sb_cc +
|
|
|
|
(tp->rcv_up - tp->rcv_nxt) - 1;
|
|
|
|
if (so->so_oobmark == 0)
|
|
|
|
so->so_state |= SS_RCVATMARK;
|
|
|
|
sohasoutofband(so);
|
|
|
|
tp->t_oobflags &= ~(TCPOOB_HAVEDATA | TCPOOB_HADDATA);
|
|
|
|
}
|
|
|
|
/*
|
|
|
|
* Remove out of band data so doesn't get presented to user.
|
|
|
|
* This can happen independent of advancing the URG pointer,
|
|
|
|
* but if two URG's are pending at once, some out-of-band
|
|
|
|
* data may creep in... ick.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if (th->th_urp <= (u_long)tlen
|
1994-05-24 10:09:53 +00:00
|
|
|
#ifdef SO_OOBINLINE
|
|
|
|
&& (so->so_options & SO_OOBINLINE) == 0
|
|
|
|
#endif
|
|
|
|
)
|
2000-01-09 19:17:30 +00:00
|
|
|
tcp_pulloutofband(so, th, m,
|
|
|
|
drop_hdrlen); /* hdr drop is delayed */
|
1994-05-24 10:09:53 +00:00
|
|
|
} else
|
|
|
|
/*
|
|
|
|
* If no out of band data is expected,
|
|
|
|
* pull receive urgent pointer along
|
|
|
|
* with the receive window.
|
|
|
|
*/
|
|
|
|
if (SEQ_GT(tp->rcv_nxt, tp->rcv_up))
|
|
|
|
tp->rcv_up = tp->rcv_nxt;
|
|
|
|
dodata: /* XXX */
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Process the segment text, merging it into the TCP sequencing queue,
|
|
|
|
* and arranging for acknowledgment of receipt if necessary.
|
|
|
|
* This process logically involves adjusting tp->rcv_wnd as data
|
|
|
|
* is presented to the user (this happens in tcp_usrreq.c,
|
|
|
|
* case PRU_RCVD). If a FIN has already been received on this
|
|
|
|
* connection then we just ignore the text.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if ((tlen || (thflags&TH_FIN)) &&
|
1994-05-24 10:09:53 +00:00
|
|
|
TCPS_HAVERCVDFIN(tp->t_state) == 0) {
|
2000-01-09 19:17:30 +00:00
|
|
|
m_adj(m, drop_hdrlen); /* delayed header drop */
|
2001-05-29 19:54:45 +00:00
|
|
|
/*
|
|
|
|
* Insert segment which inludes th into reassembly queue of tcp with
|
|
|
|
* control block tp. Return TH_FIN if reassembly now includes
|
|
|
|
* a segment with FIN. This handle 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).
|
|
|
|
*/
|
|
|
|
if (th->th_seq == tp->rcv_nxt &&
|
|
|
|
LIST_EMPTY(&tp->t_segq) &&
|
|
|
|
TCPS_HAVEESTABLISHED(tp->t_state)) {
|
|
|
|
if (DELAY_ACK(tp))
|
|
|
|
callout_reset(tp->tt_delack, tcp_delacktime,
|
|
|
|
tcp_timer_delack, tp);
|
|
|
|
else
|
|
|
|
tp->t_flags |= TF_ACKNOW;
|
|
|
|
tp->rcv_nxt += tlen;
|
|
|
|
thflags = th->th_flags & TH_FIN;
|
|
|
|
tcpstat.tcps_rcvpack++;
|
|
|
|
tcpstat.tcps_rcvbyte += tlen;
|
|
|
|
ND6_HINT(tp);
|
|
|
|
sbappend(&so->so_rcv, m);
|
|
|
|
sorwakeup(so);
|
|
|
|
} else {
|
|
|
|
thflags = tcp_reass(tp, th, &tlen, m);
|
|
|
|
tp->t_flags |= TF_ACKNOW;
|
|
|
|
}
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* Note the amount of data that peer has sent into
|
|
|
|
* our window, in order to estimate the sender's
|
|
|
|
* buffer size.
|
|
|
|
*/
|
|
|
|
len = so->so_rcv.sb_hiwat - (tp->rcv_adv - tp->rcv_nxt);
|
|
|
|
} else {
|
|
|
|
m_freem(m);
|
2000-01-09 19:17:30 +00:00
|
|
|
thflags &= ~TH_FIN;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If FIN is received ACK the FIN and let the user know
|
|
|
|
* that the connection is closing.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if (thflags & TH_FIN) {
|
1994-05-24 10:09:53 +00:00
|
|
|
if (TCPS_HAVERCVDFIN(tp->t_state) == 0) {
|
|
|
|
socantrcvmore(so);
|
1995-02-09 23:13:27 +00:00
|
|
|
/*
|
|
|
|
* If connection is half-synchronized
|
1995-07-31 10:24:22 +00:00
|
|
|
* (ie NEEDSYN flag on) then delay ACK,
|
1995-02-09 23:13:27 +00:00
|
|
|
* so it may be piggybacked when SYN is sent.
|
|
|
|
* Otherwise, since we received a FIN then no
|
1995-05-30 08:16:23 +00:00
|
|
|
* more input can be expected, send ACK now.
|
1995-02-09 23:13:27 +00:00
|
|
|
*/
|
2001-02-25 15:17:24 +00:00
|
|
|
if (DELAY_ACK(tp) && (tp->t_flags & TF_NEEDSYN))
|
1999-08-30 21:17:07 +00:00
|
|
|
callout_reset(tp->tt_delack, tcp_delacktime,
|
|
|
|
tcp_timer_delack, tp);
|
1995-05-30 08:16:23 +00:00
|
|
|
else
|
1995-02-09 23:13:27 +00:00
|
|
|
tp->t_flags |= TF_ACKNOW;
|
1994-05-24 10:09:53 +00:00
|
|
|
tp->rcv_nxt++;
|
|
|
|
}
|
|
|
|
switch (tp->t_state) {
|
|
|
|
|
|
|
|
/*
|
|
|
|
* In SYN_RECEIVED and ESTABLISHED STATES
|
|
|
|
* enter the CLOSE_WAIT state.
|
|
|
|
*/
|
|
|
|
case TCPS_SYN_RECEIVED:
|
1999-08-30 21:17:07 +00:00
|
|
|
tp->t_starttime = ticks;
|
|
|
|
/*FALLTHROUGH*/
|
1994-05-24 10:09:53 +00:00
|
|
|
case TCPS_ESTABLISHED:
|
|
|
|
tp->t_state = TCPS_CLOSE_WAIT;
|
|
|
|
break;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If still in FIN_WAIT_1 STATE FIN has not been acked so
|
|
|
|
* enter the CLOSING state.
|
|
|
|
*/
|
|
|
|
case TCPS_FIN_WAIT_1:
|
|
|
|
tp->t_state = TCPS_CLOSING;
|
|
|
|
break;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* In FIN_WAIT_2 state enter the TIME_WAIT state,
|
1995-05-30 08:16:23 +00:00
|
|
|
* starting the time-wait timer, turning off the other
|
1994-05-24 10:09:53 +00:00
|
|
|
* standard timers.
|
|
|
|
*/
|
|
|
|
case TCPS_FIN_WAIT_2:
|
|
|
|
tp->t_state = TCPS_TIME_WAIT;
|
|
|
|
tcp_canceltimers(tp);
|
1995-02-09 23:13:27 +00:00
|
|
|
/* Shorten TIME_WAIT [RFC-1644, p.28] */
|
|
|
|
if (tp->cc_recv != 0 &&
|
1999-08-30 21:17:07 +00:00
|
|
|
(ticks - tp->t_starttime) < tcp_msl) {
|
|
|
|
callout_reset(tp->tt_2msl,
|
|
|
|
tp->t_rxtcur * TCPTV_TWTRUNC,
|
|
|
|
tcp_timer_2msl, tp);
|
1995-02-09 23:13:27 +00:00
|
|
|
/* For transaction client, force ACK now. */
|
|
|
|
tp->t_flags |= TF_ACKNOW;
|
|
|
|
}
|
|
|
|
else
|
1999-08-30 21:17:07 +00:00
|
|
|
callout_reset(tp->tt_2msl, 2 * tcp_msl,
|
|
|
|
tcp_timer_2msl, tp);
|
1994-05-24 10:09:53 +00:00
|
|
|
soisdisconnected(so);
|
|
|
|
break;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* In TIME_WAIT state restart the 2 MSL time_wait timer.
|
|
|
|
*/
|
|
|
|
case TCPS_TIME_WAIT:
|
1999-08-30 21:17:07 +00:00
|
|
|
callout_reset(tp->tt_2msl, 2 * tcp_msl,
|
|
|
|
tcp_timer_2msl, tp);
|
1994-05-24 10:09:53 +00:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
1994-09-15 10:36:56 +00:00
|
|
|
#ifdef TCPDEBUG
|
1994-05-24 10:09:53 +00:00
|
|
|
if (so->so_options & SO_DEBUG)
|
2000-01-09 19:17:30 +00:00
|
|
|
tcp_trace(TA_INPUT, ostate, tp, (void *)tcp_saveipgen,
|
|
|
|
&tcp_savetcp, 0);
|
1994-09-15 10:36:56 +00:00
|
|
|
#endif
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Return any desired output.
|
|
|
|
*/
|
|
|
|
if (needoutput || (tp->t_flags & TF_ACKNOW))
|
|
|
|
(void) tcp_output(tp);
|
|
|
|
return;
|
|
|
|
|
|
|
|
dropafterack:
|
|
|
|
/*
|
|
|
|
* Generate an ACK dropping incoming segment if it occupies
|
|
|
|
* sequence space, where the ACK reflects our state.
|
1998-09-11 16:04:03 +00:00
|
|
|
*
|
|
|
|
* We can now skip the test for the RST flag since all
|
|
|
|
* paths to this code happen after packets containing
|
|
|
|
* RST have been dropped.
|
|
|
|
*
|
|
|
|
* In the SYN-RECEIVED state, don't send an ACK unless the
|
|
|
|
* segment we received passes the SYN-RECEIVED ACK test.
|
|
|
|
* If it fails send a RST. This breaks the loop in the
|
|
|
|
* "LAND" DoS attack, and also prevents an ACK storm
|
|
|
|
* between two listening ports that have been sent forged
|
|
|
|
* SYN segments, each with the source address of the other.
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if (tp->t_state == TCPS_SYN_RECEIVED && (thflags & TH_ACK) &&
|
|
|
|
(SEQ_GT(tp->snd_una, th->th_ack) ||
|
2001-02-11 07:39:51 +00:00
|
|
|
SEQ_GT(th->th_ack, tp->snd_max)) ) {
|
|
|
|
rstreason = BANDLIM_RST_OPENPORT;
|
|
|
|
goto dropwithreset;
|
|
|
|
}
|
1995-02-09 23:13:27 +00:00
|
|
|
#ifdef TCPDEBUG
|
|
|
|
if (so->so_options & SO_DEBUG)
|
2000-01-09 19:17:30 +00:00
|
|
|
tcp_trace(TA_DROP, ostate, tp, (void *)tcp_saveipgen,
|
|
|
|
&tcp_savetcp, 0);
|
1995-02-09 23:13:27 +00:00
|
|
|
#endif
|
1994-05-24 10:09:53 +00:00
|
|
|
m_freem(m);
|
|
|
|
tp->t_flags |= TF_ACKNOW;
|
|
|
|
(void) tcp_output(tp);
|
|
|
|
return;
|
|
|
|
|
|
|
|
dropwithreset:
|
|
|
|
/*
|
|
|
|
* Generate a RST, dropping incoming segment.
|
|
|
|
* Make ACK acceptable to originator of segment.
|
|
|
|
* Don't bother to respond if destination was broadcast/multicast.
|
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if ((thflags & TH_RST) || m->m_flags & (M_BCAST|M_MCAST))
|
1994-05-24 10:09:53 +00:00
|
|
|
goto drop;
|
2000-01-09 19:17:30 +00:00
|
|
|
#ifdef INET6
|
|
|
|
if (isipv6) {
|
2000-01-28 06:13:09 +00:00
|
|
|
if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) ||
|
|
|
|
IN6_IS_ADDR_MULTICAST(&ip6->ip6_src))
|
2000-01-09 19:17:30 +00:00
|
|
|
goto drop;
|
|
|
|
} else
|
|
|
|
#endif /* INET6 */
|
2000-01-28 06:13:09 +00:00
|
|
|
if (IN_MULTICAST(ntohl(ip->ip_dst.s_addr)) ||
|
|
|
|
IN_MULTICAST(ntohl(ip->ip_src.s_addr)) ||
|
|
|
|
ip->ip_src.s_addr == htonl(INADDR_BROADCAST))
|
2000-01-09 19:17:30 +00:00
|
|
|
goto drop;
|
|
|
|
/* IPv6 anycast check is done at tcp6_input() */
|
2001-02-11 07:39:51 +00:00
|
|
|
|
|
|
|
/*
|
2001-03-19 22:09:00 +00:00
|
|
|
* Perform bandwidth limiting.
|
2001-02-11 07:39:51 +00:00
|
|
|
*/
|
|
|
|
if (badport_bandlim(rstreason) < 0)
|
|
|
|
goto drop;
|
|
|
|
|
1995-02-09 23:13:27 +00:00
|
|
|
#ifdef TCPDEBUG
|
|
|
|
if (tp == 0 || (tp->t_inpcb->inp_socket->so_options & SO_DEBUG))
|
2000-01-09 19:17:30 +00:00
|
|
|
tcp_trace(TA_DROP, ostate, tp, (void *)tcp_saveipgen,
|
|
|
|
&tcp_savetcp, 0);
|
1995-02-09 23:13:27 +00:00
|
|
|
#endif
|
2000-01-09 19:17:30 +00:00
|
|
|
if (thflags & TH_ACK)
|
|
|
|
/* mtod() below is safe as long as hdr dropping is delayed */
|
|
|
|
tcp_respond(tp, mtod(m, void *), th, m, (tcp_seq)0, th->th_ack,
|
|
|
|
TH_RST);
|
1994-05-24 10:09:53 +00:00
|
|
|
else {
|
2000-01-09 19:17:30 +00:00
|
|
|
if (thflags & TH_SYN)
|
|
|
|
tlen++;
|
|
|
|
/* mtod() below is safe as long as hdr dropping is delayed */
|
|
|
|
tcp_respond(tp, mtod(m, void *), th, m, th->th_seq+tlen,
|
|
|
|
(tcp_seq)0, TH_RST|TH_ACK);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
return;
|
|
|
|
|
|
|
|
drop:
|
|
|
|
/*
|
|
|
|
* Drop space held by incoming segment and return.
|
|
|
|
*/
|
1994-09-15 10:36:56 +00:00
|
|
|
#ifdef TCPDEBUG
|
1995-02-09 23:13:27 +00:00
|
|
|
if (tp == 0 || (tp->t_inpcb->inp_socket->so_options & SO_DEBUG))
|
2000-01-09 19:17:30 +00:00
|
|
|
tcp_trace(TA_DROP, ostate, tp, (void *)tcp_saveipgen,
|
|
|
|
&tcp_savetcp, 0);
|
1994-09-15 10:36:56 +00:00
|
|
|
#endif
|
1994-05-24 10:09:53 +00:00
|
|
|
m_freem(m);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2001-11-22 04:50:44 +00:00
|
|
|
/*
|
|
|
|
* Parse TCP options and place in tcpopt.
|
|
|
|
*/
|
1995-11-14 20:34:56 +00:00
|
|
|
static void
|
2001-11-22 04:50:44 +00:00
|
|
|
tcp_dooptions(to, cp, cnt, is_syn)
|
|
|
|
struct tcpopt *to;
|
1994-05-24 10:09:53 +00:00
|
|
|
u_char *cp;
|
|
|
|
int cnt;
|
|
|
|
{
|
|
|
|
int opt, optlen;
|
|
|
|
|
2001-11-22 04:50:44 +00:00
|
|
|
to->to_flags = 0;
|
1994-05-24 10:09:53 +00:00
|
|
|
for (; cnt > 0; cnt -= optlen, cp += optlen) {
|
|
|
|
opt = cp[0];
|
|
|
|
if (opt == TCPOPT_EOL)
|
|
|
|
break;
|
|
|
|
if (opt == TCPOPT_NOP)
|
|
|
|
optlen = 1;
|
|
|
|
else {
|
2000-07-09 13:01:59 +00:00
|
|
|
if (cnt < 2)
|
|
|
|
break;
|
1994-05-24 10:09:53 +00:00
|
|
|
optlen = cp[1];
|
2000-07-09 13:01:59 +00:00
|
|
|
if (optlen < 2 || optlen > cnt)
|
1994-05-24 10:09:53 +00:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
switch (opt) {
|
|
|
|
case TCPOPT_MAXSEG:
|
|
|
|
if (optlen != TCPOLEN_MAXSEG)
|
|
|
|
continue;
|
2001-11-22 04:50:44 +00:00
|
|
|
if (!is_syn)
|
1994-05-24 10:09:53 +00:00
|
|
|
continue;
|
2001-11-22 04:50:44 +00:00
|
|
|
to->to_flags |= TOF_MSS;
|
|
|
|
bcopy((char *)cp + 2,
|
|
|
|
(char *)&to->to_mss, sizeof(to->to_mss));
|
|
|
|
NTOHS(to->to_mss);
|
1994-05-24 10:09:53 +00:00
|
|
|
break;
|
|
|
|
case TCPOPT_WINDOW:
|
|
|
|
if (optlen != TCPOLEN_WINDOW)
|
|
|
|
continue;
|
2001-11-22 04:50:44 +00:00
|
|
|
if (! is_syn)
|
1994-05-24 10:09:53 +00:00
|
|
|
continue;
|
2001-11-22 04:50:44 +00:00
|
|
|
to->to_flags |= TOF_SCALE;
|
|
|
|
to->to_requested_s_scale = min(cp[2], TCP_MAX_WINSHIFT);
|
1994-05-24 10:09:53 +00:00
|
|
|
break;
|
|
|
|
case TCPOPT_TIMESTAMP:
|
|
|
|
if (optlen != TCPOLEN_TIMESTAMP)
|
|
|
|
continue;
|
2001-11-22 04:50:44 +00:00
|
|
|
to->to_flags |= TOF_TS;
|
1995-02-09 23:13:27 +00:00
|
|
|
bcopy((char *)cp + 2,
|
|
|
|
(char *)&to->to_tsval, sizeof(to->to_tsval));
|
|
|
|
NTOHL(to->to_tsval);
|
|
|
|
bcopy((char *)cp + 6,
|
|
|
|
(char *)&to->to_tsecr, sizeof(to->to_tsecr));
|
|
|
|
NTOHL(to->to_tsecr);
|
1994-05-24 10:09:53 +00:00
|
|
|
break;
|
1995-02-09 23:13:27 +00:00
|
|
|
case TCPOPT_CC:
|
|
|
|
if (optlen != TCPOLEN_CC)
|
|
|
|
continue;
|
2001-11-22 04:50:44 +00:00
|
|
|
to->to_flags |= TOF_CC;
|
1995-02-09 23:13:27 +00:00
|
|
|
bcopy((char *)cp + 2,
|
|
|
|
(char *)&to->to_cc, sizeof(to->to_cc));
|
|
|
|
NTOHL(to->to_cc);
|
|
|
|
break;
|
|
|
|
case TCPOPT_CCNEW:
|
|
|
|
if (optlen != TCPOLEN_CC)
|
|
|
|
continue;
|
2001-11-22 04:50:44 +00:00
|
|
|
if (!is_syn)
|
1995-02-09 23:13:27 +00:00
|
|
|
continue;
|
2001-11-22 04:50:44 +00:00
|
|
|
to->to_flags |= TOF_CCNEW;
|
1995-02-09 23:13:27 +00:00
|
|
|
bcopy((char *)cp + 2,
|
|
|
|
(char *)&to->to_cc, sizeof(to->to_cc));
|
|
|
|
NTOHL(to->to_cc);
|
|
|
|
break;
|
|
|
|
case TCPOPT_CCECHO:
|
|
|
|
if (optlen != TCPOLEN_CC)
|
|
|
|
continue;
|
2001-11-22 04:50:44 +00:00
|
|
|
if (!is_syn)
|
1995-02-09 23:13:27 +00:00
|
|
|
continue;
|
2001-11-22 04:50:44 +00:00
|
|
|
to->to_flags |= TOF_CCECHO;
|
1995-02-09 23:13:27 +00:00
|
|
|
bcopy((char *)cp + 2,
|
|
|
|
(char *)&to->to_ccecho, sizeof(to->to_ccecho));
|
|
|
|
NTOHL(to->to_ccecho);
|
|
|
|
break;
|
2001-11-22 04:50:44 +00:00
|
|
|
default:
|
|
|
|
continue;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Pull out of band byte out of a segment so
|
|
|
|
* it doesn't appear in the user's data queue.
|
|
|
|
* It is still reflected in the segment length for
|
|
|
|
* sequencing purposes.
|
|
|
|
*/
|
1995-11-14 20:34:56 +00:00
|
|
|
static void
|
2000-01-09 19:17:30 +00:00
|
|
|
tcp_pulloutofband(so, th, m, off)
|
1994-05-24 10:09:53 +00:00
|
|
|
struct socket *so;
|
2000-01-09 19:17:30 +00:00
|
|
|
struct tcphdr *th;
|
1994-05-24 10:09:53 +00:00
|
|
|
register struct mbuf *m;
|
2000-01-09 19:17:30 +00:00
|
|
|
int off; /* delayed to be droped hdrlen */
|
1994-05-24 10:09:53 +00:00
|
|
|
{
|
2000-01-09 19:17:30 +00:00
|
|
|
int cnt = off + th->th_urp - 1;
|
1995-05-30 08:16:23 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
while (cnt >= 0) {
|
|
|
|
if (m->m_len > cnt) {
|
|
|
|
char *cp = mtod(m, caddr_t) + cnt;
|
|
|
|
struct tcpcb *tp = sototcpcb(so);
|
|
|
|
|
|
|
|
tp->t_iobc = *cp;
|
|
|
|
tp->t_oobflags |= TCPOOB_HAVEDATA;
|
|
|
|
bcopy(cp+1, cp, (unsigned)(m->m_len - cnt - 1));
|
|
|
|
m->m_len--;
|
2000-01-25 01:26:47 +00:00
|
|
|
if (m->m_flags & M_PKTHDR)
|
|
|
|
m->m_pkthdr.len--;
|
1994-05-24 10:09:53 +00:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
cnt -= m->m_len;
|
|
|
|
m = m->m_next;
|
|
|
|
if (m == 0)
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
panic("tcp_pulloutofband");
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Collect new round-trip time estimate
|
|
|
|
* and update averages and current timeout.
|
|
|
|
*/
|
1995-11-14 20:34:56 +00:00
|
|
|
static void
|
1994-05-24 10:09:53 +00:00
|
|
|
tcp_xmit_timer(tp, rtt)
|
|
|
|
register struct tcpcb *tp;
|
1999-08-30 21:17:07 +00:00
|
|
|
int rtt;
|
1994-05-24 10:09:53 +00:00
|
|
|
{
|
1996-03-22 18:09:21 +00:00
|
|
|
register int delta;
|
|
|
|
|
|
|
|
tcpstat.tcps_rttupdated++;
|
|
|
|
tp->t_rttupdated++;
|
|
|
|
if (tp->t_srtt != 0) {
|
|
|
|
/*
|
|
|
|
* srtt is stored as fixed point with 5 bits after the
|
|
|
|
* binary point (i.e., scaled by 8). The following magic
|
|
|
|
* is equivalent to the smoothing algorithm in rfc793 with
|
|
|
|
* an alpha of .875 (srtt = rtt/8 + srtt*7/8 in fixed
|
|
|
|
* point). Adjust rtt to origin 0.
|
|
|
|
*/
|
|
|
|
delta = ((rtt - 1) << TCP_DELTA_SHIFT)
|
|
|
|
- (tp->t_srtt >> (TCP_RTT_SHIFT - TCP_DELTA_SHIFT));
|
|
|
|
|
|
|
|
if ((tp->t_srtt += delta) <= 0)
|
|
|
|
tp->t_srtt = 1;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* We accumulate a smoothed rtt variance (actually, a
|
|
|
|
* smoothed mean difference), then set the retransmit
|
|
|
|
* timer to smoothed rtt + 4 times the smoothed variance.
|
|
|
|
* rttvar is stored as fixed point with 4 bits after the
|
|
|
|
* binary point (scaled by 16). The following is
|
|
|
|
* equivalent to rfc793 smoothing with an alpha of .75
|
|
|
|
* (rttvar = rttvar*3/4 + |delta| / 4). This replaces
|
|
|
|
* rfc793's wired-in beta.
|
|
|
|
*/
|
|
|
|
if (delta < 0)
|
|
|
|
delta = -delta;
|
|
|
|
delta -= tp->t_rttvar >> (TCP_RTTVAR_SHIFT - TCP_DELTA_SHIFT);
|
|
|
|
if ((tp->t_rttvar += delta) <= 0)
|
|
|
|
tp->t_rttvar = 1;
|
|
|
|
} else {
|
|
|
|
/*
|
|
|
|
* No rtt measurement yet - use the unsmoothed rtt.
|
|
|
|
* Set the variance to half the rtt (so our first
|
|
|
|
* retransmit happens at 3*rtt).
|
|
|
|
*/
|
|
|
|
tp->t_srtt = rtt << TCP_RTT_SHIFT;
|
|
|
|
tp->t_rttvar = rtt << (TCP_RTTVAR_SHIFT - 1);
|
|
|
|
}
|
1999-08-30 21:17:07 +00:00
|
|
|
tp->t_rtttime = 0;
|
1994-05-24 10:09:53 +00:00
|
|
|
tp->t_rxtshift = 0;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* the retransmit should happen at rtt + 4 * rttvar.
|
|
|
|
* Because of the way we do the smoothing, srtt and rttvar
|
|
|
|
* will each average +1/2 tick of bias. When we compute
|
|
|
|
* the retransmit timer, we want 1/2 tick of rounding and
|
|
|
|
* 1 extra tick because of +-1/2 tick uncertainty in the
|
|
|
|
* firing of the timer. The bias will give us exactly the
|
|
|
|
* 1.5 tick we need. But, because the bias is
|
|
|
|
* statistical, we have to test that we don't drop below
|
|
|
|
* the minimum feasible timer (which is 2 ticks).
|
|
|
|
*/
|
1996-03-22 18:09:21 +00:00
|
|
|
TCPT_RANGESET(tp->t_rxtcur, TCP_REXMTVAL(tp),
|
1996-03-25 20:13:21 +00:00
|
|
|
max(tp->t_rttmin, rtt + 2), TCPTV_REXMTMAX);
|
1995-05-30 08:16:23 +00:00
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* We received an ack for a packet that wasn't retransmitted;
|
|
|
|
* it is probably safe to discard any error indications we've
|
|
|
|
* received recently. This isn't quite right, but close enough
|
|
|
|
* for now (a route might have failed after we sent a segment,
|
|
|
|
* and the return path might not be symmetrical).
|
|
|
|
*/
|
|
|
|
tp->t_softerror = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Determine a reasonable value for maxseg size.
|
|
|
|
* If the route is known, check route for mtu.
|
|
|
|
* If none, use an mss that can be handled on the outgoing
|
|
|
|
* interface without forcing IP to fragment; if bigger than
|
|
|
|
* an mbuf cluster (MCLBYTES), round down to nearest multiple of MCLBYTES
|
|
|
|
* to utilize large mbufs. If no route is found, route has no mtu,
|
|
|
|
* or the destination isn't local, use a default, hopefully conservative
|
|
|
|
* size (usually 512 or the default IP max size, but no more than the mtu
|
|
|
|
* of the interface), as we can't discover anything about intervening
|
|
|
|
* gateways or networks. We also initialize the congestion/slow start
|
|
|
|
* window to be a single segment if the destination isn't local.
|
|
|
|
* While looking at the routing entry, we also initialize other path-dependent
|
|
|
|
* parameters from pre-set or cached values in the routing entry.
|
1995-02-09 23:13:27 +00:00
|
|
|
*
|
|
|
|
* Also take into account the space needed for options that we
|
|
|
|
* send regularly. Make maxseg shorter by that amount to assure
|
|
|
|
* that we can send maxseg amount of data even when the options
|
|
|
|
* are present. Store the upper limit of the length of options plus
|
|
|
|
* data in maxopd.
|
|
|
|
*
|
|
|
|
* NOTE that this routine is only called when we process an incoming
|
|
|
|
* segment, for outgoing segments only tcp_mssopt is called.
|
|
|
|
*
|
|
|
|
* In case of T/TCP, we call this routine during implicit connection
|
|
|
|
* setup as well (offer = -1), to initialize maxseg from the cached
|
|
|
|
* MSS of our peer.
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
1995-02-09 23:13:27 +00:00
|
|
|
void
|
1994-05-24 10:09:53 +00:00
|
|
|
tcp_mss(tp, offer)
|
1995-02-09 23:13:27 +00:00
|
|
|
struct tcpcb *tp;
|
|
|
|
int offer;
|
1994-05-24 10:09:53 +00:00
|
|
|
{
|
|
|
|
register struct rtentry *rt;
|
|
|
|
struct ifnet *ifp;
|
|
|
|
register int rtt, mss;
|
|
|
|
u_long bufsize;
|
|
|
|
struct inpcb *inp;
|
|
|
|
struct socket *so;
|
1995-02-09 23:13:27 +00:00
|
|
|
struct rmxp_tao *taop;
|
|
|
|
int origoffer = offer;
|
2000-01-09 19:17:30 +00:00
|
|
|
#ifdef INET6
|
|
|
|
int isipv6;
|
|
|
|
int min_protoh;
|
|
|
|
#endif
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
inp = tp->t_inpcb;
|
2000-01-09 19:17:30 +00:00
|
|
|
#ifdef INET6
|
|
|
|
isipv6 = ((inp->inp_vflag & INP_IPV6) != 0) ? 1 : 0;
|
|
|
|
min_protoh = isipv6 ? sizeof (struct ip6_hdr) + sizeof (struct tcphdr)
|
|
|
|
: sizeof (struct tcpiphdr);
|
|
|
|
#else
|
|
|
|
#define min_protoh (sizeof (struct tcpiphdr))
|
|
|
|
#endif
|
|
|
|
#ifdef INET6
|
|
|
|
if (isipv6)
|
2001-11-22 04:50:44 +00:00
|
|
|
rt = tcp_rtlookup6(&inp->inp_inc);
|
2000-01-09 19:17:30 +00:00
|
|
|
else
|
|
|
|
#endif
|
2001-11-22 04:50:44 +00:00
|
|
|
rt = tcp_rtlookup(&inp->inp_inc);
|
2000-01-09 19:17:30 +00:00
|
|
|
if (rt == NULL) {
|
|
|
|
tp->t_maxopd = tp->t_maxseg =
|
|
|
|
#ifdef INET6
|
|
|
|
isipv6 ? tcp_v6mssdflt :
|
|
|
|
#endif /* INET6 */
|
|
|
|
tcp_mssdflt;
|
1995-02-09 23:13:27 +00:00
|
|
|
return;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
ifp = rt->rt_ifp;
|
|
|
|
so = inp->inp_socket;
|
|
|
|
|
1995-02-09 23:13:27 +00:00
|
|
|
taop = rmx_taop(rt->rt_rmx);
|
|
|
|
/*
|
|
|
|
* Offer == -1 means that we didn't receive SYN yet,
|
|
|
|
* use cached value in that case;
|
|
|
|
*/
|
|
|
|
if (offer == -1)
|
|
|
|
offer = taop->tao_mssopt;
|
|
|
|
/*
|
|
|
|
* Offer == 0 means that there was no MSS on the SYN segment,
|
|
|
|
* in this case we use tcp_mssdflt.
|
|
|
|
*/
|
|
|
|
if (offer == 0)
|
2000-01-09 19:17:30 +00:00
|
|
|
offer =
|
|
|
|
#ifdef INET6
|
|
|
|
isipv6 ? tcp_v6mssdflt :
|
|
|
|
#endif /* INET6 */
|
|
|
|
tcp_mssdflt;
|
1995-02-09 23:13:27 +00:00
|
|
|
else
|
|
|
|
/*
|
|
|
|
* Sanity check: make sure that maxopd will be large
|
|
|
|
* enough to allow some data on segments even is the
|
|
|
|
* all the option space is used (40bytes). Otherwise
|
|
|
|
* funny things may happen in tcp_output.
|
|
|
|
*/
|
|
|
|
offer = max(offer, 64);
|
|
|
|
taop->tao_mssopt = offer;
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
/*
|
|
|
|
* While we're here, check if there's an initial rtt
|
|
|
|
* or rttvar. Convert from the route-table units
|
|
|
|
* to scaled multiples of the slow timeout timer.
|
|
|
|
*/
|
|
|
|
if (tp->t_srtt == 0 && (rtt = rt->rt_rmx.rmx_rtt)) {
|
|
|
|
/*
|
1995-02-09 23:13:27 +00:00
|
|
|
* XXX the lock bit for RTT indicates that the value
|
1994-05-24 10:09:53 +00:00
|
|
|
* is also a minimum value; this is subject to time.
|
|
|
|
*/
|
|
|
|
if (rt->rt_rmx.rmx_locks & RTV_RTT)
|
1999-08-30 21:17:07 +00:00
|
|
|
tp->t_rttmin = rtt / (RTM_RTTUNIT / hz);
|
|
|
|
tp->t_srtt = rtt / (RTM_RTTUNIT / (hz * TCP_RTT_SCALE));
|
1995-07-10 15:39:16 +00:00
|
|
|
tcpstat.tcps_usedrtt++;
|
|
|
|
if (rt->rt_rmx.rmx_rttvar) {
|
1994-05-24 10:09:53 +00:00
|
|
|
tp->t_rttvar = rt->rt_rmx.rmx_rttvar /
|
1999-08-30 21:17:07 +00:00
|
|
|
(RTM_RTTUNIT / (hz * TCP_RTTVAR_SCALE));
|
1995-07-10 15:39:16 +00:00
|
|
|
tcpstat.tcps_usedrttvar++;
|
|
|
|
} else {
|
1994-05-24 10:09:53 +00:00
|
|
|
/* default variation is +- 1 rtt */
|
|
|
|
tp->t_rttvar =
|
|
|
|
tp->t_srtt * TCP_RTTVAR_SCALE / TCP_RTT_SCALE;
|
1995-07-10 15:39:16 +00:00
|
|
|
}
|
1994-05-24 10:09:53 +00:00
|
|
|
TCPT_RANGESET(tp->t_rxtcur,
|
1999-08-30 21:17:07 +00:00
|
|
|
((tp->t_srtt >> 2) + tp->t_rttvar) >> 1,
|
|
|
|
tp->t_rttmin, TCPTV_REXMTMAX);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
|
|
|
/*
|
|
|
|
* if there's an mtu associated with the route, use it
|
2000-01-09 19:17:30 +00:00
|
|
|
* else, use the link mtu.
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
|
|
|
if (rt->rt_rmx.rmx_mtu)
|
2000-01-09 19:17:30 +00:00
|
|
|
mss = rt->rt_rmx.rmx_mtu - min_protoh;
|
1994-05-24 10:09:53 +00:00
|
|
|
else
|
|
|
|
{
|
2000-01-09 19:17:30 +00:00
|
|
|
mss =
|
|
|
|
#ifdef INET6
|
|
|
|
(isipv6 ? nd_ifinfo[rt->rt_ifp->if_index].linkmtu :
|
|
|
|
#endif
|
|
|
|
ifp->if_mtu
|
|
|
|
#ifdef INET6
|
|
|
|
)
|
|
|
|
#endif
|
|
|
|
- min_protoh;
|
|
|
|
#ifdef INET6
|
|
|
|
if (isipv6) {
|
|
|
|
if (!in6_localaddr(&inp->in6p_faddr))
|
|
|
|
mss = min(mss, tcp_v6mssdflt);
|
|
|
|
} else
|
|
|
|
#endif
|
1995-02-09 23:13:27 +00:00
|
|
|
if (!in_localaddr(inp->inp_faddr))
|
|
|
|
mss = min(mss, tcp_mssdflt);
|
|
|
|
}
|
|
|
|
mss = min(mss, offer);
|
|
|
|
/*
|
|
|
|
* maxopd stores the maximum length of data AND options
|
|
|
|
* in a segment; maxseg is the amount of data in a normal
|
|
|
|
* segment. We need to store this value (maxopd) apart
|
|
|
|
* from maxseg, because now every segment carries options
|
|
|
|
* and thus we normally have somewhat less data in segments.
|
|
|
|
*/
|
|
|
|
tp->t_maxopd = mss;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* In case of T/TCP, origoffer==-1 indicates, that no segments
|
|
|
|
* were received yet. In this case we just guess, otherwise
|
|
|
|
* we do the same as before T/TCP.
|
|
|
|
*/
|
|
|
|
if ((tp->t_flags & (TF_REQ_TSTMP|TF_NOOPT)) == TF_REQ_TSTMP &&
|
|
|
|
(origoffer == -1 ||
|
|
|
|
(tp->t_flags & TF_RCVD_TSTMP) == TF_RCVD_TSTMP))
|
|
|
|
mss -= TCPOLEN_TSTAMP_APPA;
|
|
|
|
if ((tp->t_flags & (TF_REQ_CC|TF_NOOPT)) == TF_REQ_CC &&
|
|
|
|
(origoffer == -1 ||
|
|
|
|
(tp->t_flags & TF_RCVD_CC) == TF_RCVD_CC))
|
|
|
|
mss -= TCPOLEN_CC_APPA;
|
|
|
|
|
1994-05-24 10:09:53 +00:00
|
|
|
#if (MCLBYTES & (MCLBYTES - 1)) == 0
|
|
|
|
if (mss > MCLBYTES)
|
|
|
|
mss &= ~(MCLBYTES-1);
|
|
|
|
#else
|
|
|
|
if (mss > MCLBYTES)
|
|
|
|
mss = mss / MCLBYTES * MCLBYTES;
|
|
|
|
#endif
|
|
|
|
/*
|
1995-02-09 23:13:27 +00:00
|
|
|
* If there's a pipesize, change the socket buffer
|
|
|
|
* to that size. Make the socket buffers an integral
|
|
|
|
* number of mss units; if the mss is larger than
|
|
|
|
* the socket buffer, decrease the mss.
|
1994-05-24 10:09:53 +00:00
|
|
|
*/
|
|
|
|
#ifdef RTV_SPIPE
|
1995-02-09 23:13:27 +00:00
|
|
|
if ((bufsize = rt->rt_rmx.rmx_sendpipe) == 0)
|
1994-05-24 10:09:53 +00:00
|
|
|
#endif
|
1995-02-09 23:13:27 +00:00
|
|
|
bufsize = so->so_snd.sb_hiwat;
|
|
|
|
if (bufsize < mss)
|
|
|
|
mss = bufsize;
|
|
|
|
else {
|
|
|
|
bufsize = roundup(bufsize, mss);
|
|
|
|
if (bufsize > sb_max)
|
|
|
|
bufsize = sb_max;
|
1999-10-09 20:42:17 +00:00
|
|
|
(void)sbreserve(&so->so_snd, bufsize, so, NULL);
|
1995-02-09 23:13:27 +00:00
|
|
|
}
|
|
|
|
tp->t_maxseg = mss;
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
#ifdef RTV_RPIPE
|
1995-02-09 23:13:27 +00:00
|
|
|
if ((bufsize = rt->rt_rmx.rmx_recvpipe) == 0)
|
1994-05-24 10:09:53 +00:00
|
|
|
#endif
|
1995-02-09 23:13:27 +00:00
|
|
|
bufsize = so->so_rcv.sb_hiwat;
|
|
|
|
if (bufsize > mss) {
|
|
|
|
bufsize = roundup(bufsize, mss);
|
|
|
|
if (bufsize > sb_max)
|
|
|
|
bufsize = sb_max;
|
1999-10-09 20:42:17 +00:00
|
|
|
(void)sbreserve(&so->so_rcv, bufsize, so, NULL);
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
1999-08-30 21:17:07 +00:00
|
|
|
|
1995-02-09 23:13:27 +00:00
|
|
|
/*
|
1999-08-30 21:17:07 +00:00
|
|
|
* Set the slow-start flight size depending on whether this
|
|
|
|
* is a local network or not.
|
1995-02-09 23:13:27 +00:00
|
|
|
*/
|
2000-01-09 19:17:30 +00:00
|
|
|
if (
|
|
|
|
#ifdef INET6
|
|
|
|
(isipv6 && in6_localaddr(&inp->in6p_faddr)) ||
|
|
|
|
(!isipv6 &&
|
|
|
|
#endif
|
|
|
|
in_localaddr(inp->inp_faddr)
|
|
|
|
#ifdef INET6
|
|
|
|
)
|
|
|
|
#endif
|
|
|
|
)
|
1999-08-30 21:17:07 +00:00
|
|
|
tp->snd_cwnd = mss * ss_fltsz_local;
|
|
|
|
else
|
|
|
|
tp->snd_cwnd = mss * ss_fltsz;
|
1994-05-24 10:09:53 +00:00
|
|
|
|
|
|
|
if (rt->rt_rmx.rmx_ssthresh) {
|
|
|
|
/*
|
|
|
|
* There's some sort of gateway or interface
|
|
|
|
* buffer limit on the path. Use this to set
|
|
|
|
* the slow start threshhold, but set the
|
|
|
|
* threshold to no less than 2*mss.
|
|
|
|
*/
|
|
|
|
tp->snd_ssthresh = max(2 * mss, rt->rt_rmx.rmx_ssthresh);
|
1995-07-10 15:39:16 +00:00
|
|
|
tcpstat.tcps_usedssthresh++;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
1995-02-09 23:13:27 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Determine the MSS option to send on an outgoing SYN.
|
|
|
|
*/
|
|
|
|
int
|
|
|
|
tcp_mssopt(tp)
|
|
|
|
struct tcpcb *tp;
|
|
|
|
{
|
|
|
|
struct rtentry *rt;
|
2000-01-09 19:17:30 +00:00
|
|
|
#ifdef INET6
|
|
|
|
int isipv6;
|
|
|
|
int min_protoh;
|
|
|
|
#endif
|
1995-02-09 23:13:27 +00:00
|
|
|
|
2000-01-09 19:17:30 +00:00
|
|
|
#ifdef INET6
|
|
|
|
isipv6 = ((tp->t_inpcb->inp_vflag & INP_IPV6) != 0) ? 1 : 0;
|
|
|
|
min_protoh = isipv6 ? sizeof (struct ip6_hdr) + sizeof (struct tcphdr)
|
|
|
|
: sizeof (struct tcpiphdr);
|
|
|
|
#else
|
|
|
|
#define min_protoh (sizeof (struct tcpiphdr))
|
|
|
|
#endif
|
|
|
|
#ifdef INET6
|
|
|
|
if (isipv6)
|
2001-11-22 04:50:44 +00:00
|
|
|
rt = tcp_rtlookup6(&tp->t_inpcb->inp_inc);
|
2000-01-09 19:17:30 +00:00
|
|
|
else
|
|
|
|
#endif /* INET6 */
|
2001-11-22 04:50:44 +00:00
|
|
|
rt = tcp_rtlookup(&tp->t_inpcb->inp_inc);
|
1995-02-09 23:13:27 +00:00
|
|
|
if (rt == NULL)
|
2000-01-09 19:17:30 +00:00
|
|
|
return
|
|
|
|
#ifdef INET6
|
|
|
|
isipv6 ? tcp_v6mssdflt :
|
|
|
|
#endif /* INET6 */
|
|
|
|
tcp_mssdflt;
|
1995-02-09 23:13:27 +00:00
|
|
|
|
2000-01-09 19:17:30 +00:00
|
|
|
return rt->rt_ifp->if_mtu - min_protoh;
|
1994-05-24 10:09:53 +00:00
|
|
|
}
|
2000-05-06 03:31:09 +00:00
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Checks for partial ack. If partial ack arrives, force the retransmission
|
|
|
|
* of the next unacknowledged segment, do not clear tp->t_dupacks, and return
|
|
|
|
* 1. By setting snd_nxt to ti_ack, this forces retransmission timer to
|
|
|
|
* be started again. If the ack advances at least to tp->snd_recover, return 0.
|
|
|
|
*/
|
|
|
|
static int
|
|
|
|
tcp_newreno(tp, th)
|
|
|
|
struct tcpcb *tp;
|
|
|
|
struct tcphdr *th;
|
|
|
|
{
|
|
|
|
if (SEQ_LT(th->th_ack, tp->snd_recover)) {
|
|
|
|
tcp_seq onxt = tp->snd_nxt;
|
|
|
|
u_long ocwnd = tp->snd_cwnd;
|
|
|
|
|
|
|
|
callout_stop(tp->tt_rexmt);
|
|
|
|
tp->t_rtttime = 0;
|
|
|
|
tp->snd_nxt = th->th_ack;
|
2000-05-16 03:13:59 +00:00
|
|
|
/*
|
|
|
|
* Set snd_cwnd to one segment beyond acknowledged offset
|
|
|
|
* (tp->snd_una has not yet been updated when this function
|
|
|
|
* is called)
|
|
|
|
*/
|
|
|
|
tp->snd_cwnd = tp->t_maxseg + (th->th_ack - tp->snd_una);
|
2000-05-06 03:31:09 +00:00
|
|
|
(void) tcp_output(tp);
|
|
|
|
tp->snd_cwnd = ocwnd;
|
|
|
|
if (SEQ_GT(onxt, tp->snd_nxt))
|
|
|
|
tp->snd_nxt = onxt;
|
|
|
|
/*
|
|
|
|
* Partial window deflation. Relies on fact that tp->snd_una
|
|
|
|
* not updated yet.
|
|
|
|
*/
|
|
|
|
tp->snd_cwnd -= (th->th_ack - tp->snd_una - tp->t_maxseg);
|
|
|
|
return (1);
|
|
|
|
}
|
|
|
|
return (0);
|
|
|
|
}
|