BSD 4_4_Lite2 release

[unix-history] / usr / src / sys / netinet / tcp_subr.c

diff --git a/usr/src/sys/netinet/tcp_subr.c b/usr/src/sys/netinet/tcp_subr.c
index 6e77446..9803ec9 100644 (file)
--- a/usr/src/sys/netinet/tcp_subr.c
+++ b/usr/src/sys/netinet/tcp_subr.c
@@ -1,36 +1,84 @@
-/*     tcp_subr.c      4.11    81/12/21        */
-
-#include "../h/param.h"
-#include "../h/systm.h"
-#include "../h/mbuf.h"
-#include "../h/socket.h"
-#include "../h/socketvar.h"
-#include "../h/protosw.h"
-#include "../net/in.h"
-#include "../net/in_pcb.h"
-#include "../net/in_systm.h"
-#include "../net/if.h"
-#include "../net/ip.h"
-#include "../net/ip_var.h"
-#include "../net/tcp.h"
-#include "../net/tcp_fsm.h"
-#include "../net/tcp_seq.h"
-#include "../net/tcp_timer.h"
-#include "../net/tcp_var.h"
-#include "../net/tcpip.h"
-#include "../errno.h"
+/*
+ * Copyright (c) 1982, 1986, 1988, 1990, 1993, 1995
+ *     The Regents of the University of California.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. All advertising materials mentioning features or use of this software
+ *    must display the following acknowledgement:
+ *     This product includes software developed by the University of
+ *     California, Berkeley and its contributors.
+ * 4. Neither the name of the University nor the names of its contributors
+ *    may be used to endorse or promote products derived from this software
+ *    without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ *     @(#)tcp_subr.c  8.2 (Berkeley) 5/24/95
+ */
+
+#include <sys/param.h>
+#include <sys/proc.h>
+#include <sys/systm.h>
+#include <sys/malloc.h>
+#include <sys/mbuf.h>
+#include <sys/socket.h>
+#include <sys/socketvar.h>
+#include <sys/protosw.h>
+#include <sys/errno.h>
+
+#include <net/route.h>
+#include <net/if.h>
+
+#include <netinet/in.h>
+#include <netinet/in_systm.h>
+#include <netinet/ip.h>
+#include <netinet/in_pcb.h>
+#include <netinet/ip_var.h>
+#include <netinet/ip_icmp.h>
+#include <netinet/tcp.h>
+#include <netinet/tcp_fsm.h>
+#include <netinet/tcp_seq.h>
+#include <netinet/tcp_timer.h>
+#include <netinet/tcp_var.h>
+#include <netinet/tcpip.h>
+
+/* patchable/settable parameters for tcp */
+int    tcp_mssdflt = TCP_MSS;
+int    tcp_rttdflt = TCPTV_SRTTDFLT / PR_SLOWHZ;
+int    tcp_do_rfc1323 = 1;
+
+extern struct inpcb *tcp_last_inpcb;

 
 /*
  * Tcp initialization
  */
 
 /*
  * Tcp initialization
  */

+void

 tcp_init()
 {
 
 tcp_init()
 {
 

-COUNT(TCP_INIT);
-       tcp_iss = 1;            /* wrong */
+       tcp_iss = random();     /* wrong, but better than a constant */

        tcb.inp_next = tcb.inp_prev = &tcb;
        tcb.inp_next = tcb.inp_prev = &tcb;

-       tcp_alpha = TCP_ALPHA;
-       tcp_beta = TCP_BETA;
+       if (max_protohdr < sizeof(struct tcpiphdr))
+               max_protohdr = sizeof(struct tcpiphdr);
+       if (max_linkhdr + sizeof(struct tcpiphdr) > MHLEN)
+               panic("tcp_init");

 }
 
 /*
 }
 
 /*


@@ -47,13 +95,13 @@ tcp_template(tp)
        register struct mbuf *m;
        register struct tcpiphdr *n;
 
        register struct mbuf *m;
        register struct tcpiphdr *n;
 

-COUNT(TCP_TEMPLATE);
-       m = m_get(1);
-       if (m == 0)
-               return (0);
-       m->m_off = MMAXOFF - sizeof (struct tcpiphdr);
-       m->m_len = sizeof (struct tcpiphdr);
-       n = mtod(m, struct tcpiphdr *);
+       if ((n = tp->t_template) == 0) {
+               m = m_get(M_DONTWAIT, MT_HEADER);
+               if (m == NULL)
+                       return (0);
+               m->m_len = sizeof (struct tcpiphdr);
+               n = mtod(m, struct tcpiphdr *);
+       }

        n->ti_next = n->ti_prev = 0;
        n->ti_x1 = 0;
        n->ti_pr = IPPROTO_TCP;
        n->ti_next = n->ti_prev = 0;
        n->ti_x1 = 0;
        n->ti_pr = IPPROTO_TCP;


@@ -75,7 +123,7 @@ COUNT(TCP_TEMPLATE);
 
 /*
  * Send a single message to the TCP at address specified by
 
 /*
  * Send a single message to the TCP at address specified by

- * the given TCP/IP header.  If flags==0, then we make a copy
+ * the given TCP/IP header.  If m == 0, then we make a copy

  * of the tcpiphdr at ti and send directly to the addressed host.
  * This is used to force keep alive messages out using the TCP
  * template for a connection tp->t_template.  If flags are given
  * of the tcpiphdr at ti and send directly to the addressed host.
  * This is used to force keep alive messages out using the TCP
  * template for a connection tp->t_template.  If flags are given


@@ -86,52 +134,68 @@ COUNT(TCP_TEMPLATE);
  * In any case the ack and sequence number of the transmitted
  * segment are as specified by the parameters.
  */
  * In any case the ack and sequence number of the transmitted
  * segment are as specified by the parameters.
  */

-tcp_respond(ti, ack, seq, flags)
+void
+tcp_respond(tp, ti, m, ack, seq, flags)
+       struct tcpcb *tp;

        register struct tcpiphdr *ti;
        register struct tcpiphdr *ti;

+       register struct mbuf *m;

        tcp_seq ack, seq;
        int flags;
 {
        tcp_seq ack, seq;
        int flags;
 {

-       struct mbuf *m;
+       register int tlen;
+       int win = 0;
+       struct route *ro = 0;

 
 

-COUNT(TCP_RESPOND);
-printf("tcp_respond\n");
-       if (flags == 0) {
-               m = m_get(0);
-               if (m == 0)
+       if (tp) {
+               win = sbspace(&tp->t_inpcb->inp_socket->so_rcv);
+               ro = &tp->t_inpcb->inp_route;
+       }
+       if (m == 0) {
+               m = m_gethdr(M_DONTWAIT, MT_HEADER);
+               if (m == NULL)

                        return;
                        return;

-               m->m_off = MMINOFF;
-               m->m_len = sizeof (struct tcpiphdr);
+#ifdef TCP_COMPAT_42
+               tlen = 1;
+#else
+               tlen = 0;
+#endif
+               m->m_data += max_linkhdr;

                *mtod(m, struct tcpiphdr *) = *ti;
                ti = mtod(m, struct tcpiphdr *);
                flags = TH_ACK;
        } else {
                *mtod(m, struct tcpiphdr *) = *ti;
                ti = mtod(m, struct tcpiphdr *);
                flags = TH_ACK;
        } else {

-               m = dtom(ti);

                m_freem(m->m_next);
                m->m_next = 0;
                m_freem(m->m_next);
                m->m_next = 0;

+               m->m_data = (caddr_t)ti;

                m->m_len = sizeof (struct tcpiphdr);
                m->m_len = sizeof (struct tcpiphdr);

+               tlen = 0;

 #define xchg(a,b,type) { type t; t=a; a=b; b=t; }
                xchg(ti->ti_dst.s_addr, ti->ti_src.s_addr, u_long);
                xchg(ti->ti_dport, ti->ti_sport, u_short);
 #undef xchg
        }
 #define xchg(a,b,type) { type t; t=a; a=b; b=t; }
                xchg(ti->ti_dst.s_addr, ti->ti_src.s_addr, u_long);
                xchg(ti->ti_dport, ti->ti_sport, u_short);
 #undef xchg
        }

+       ti->ti_len = htons((u_short)(sizeof (struct tcphdr) + tlen));
+       tlen += sizeof (struct tcpiphdr);
+       m->m_len = tlen;
+       m->m_pkthdr.len = tlen;
+       m->m_pkthdr.rcvif = (struct ifnet *) 0;

        ti->ti_next = ti->ti_prev = 0;
        ti->ti_x1 = 0;
        ti->ti_next = ti->ti_prev = 0;
        ti->ti_x1 = 0;

-       ti->ti_len = sizeof (struct tcphdr);
-       ti->ti_seq = seq;
-       ti->ti_ack = ack;
-#if vax
-       ti->ti_len = htons(ti->ti_len);
-       ti->ti_seq = htonl(ti->ti_seq);
-       ti->ti_ack = htonl(ti->ti_ack);
-#endif
+       ti->ti_seq = htonl(seq);
+       ti->ti_ack = htonl(ack);

        ti->ti_x2 = 0;
        ti->ti_off = sizeof (struct tcphdr) >> 2;
        ti->ti_flags = flags;
        ti->ti_x2 = 0;
        ti->ti_off = sizeof (struct tcphdr) >> 2;
        ti->ti_flags = flags;

-       ti->ti_win = ti->ti_urp = 0;
-       ti->ti_sum = in_cksum(m, sizeof(struct tcpiphdr));
-       ((struct ip *)ti)->ip_len = sizeof(struct tcpiphdr);
-       ((struct ip *)ti)->ip_ttl = TCP_TTL;
-       (void) ip_output(m, (struct mbuf *)0);
+       if (tp)
+               ti->ti_win = htons((u_short) (win >> tp->rcv_scale));
+       else
+               ti->ti_win = htons((u_short)win);
+       ti->ti_urp = 0;
+       ti->ti_sum = 0;
+       ti->ti_sum = in_cksum(m, tlen);
+       ((struct ip *)ti)->ip_len = tlen;
+       ((struct ip *)ti)->ip_ttl = ip_defttl;
+       (void) ip_output(m, NULL, ro, 0, NULL);

 }
 
 /*
 }
 
 /*


@@ -143,16 +207,31 @@ struct tcpcb *
 tcp_newtcpcb(inp)
        struct inpcb *inp;
 {
 tcp_newtcpcb(inp)
        struct inpcb *inp;
 {

-       struct mbuf *m = m_getclr(0);

        register struct tcpcb *tp;
        register struct tcpcb *tp;

-COUNT(TCP_NEWTCPCB);

 
 

-       if (m == 0)
-               return (0);
-       tp = mtod(m, struct tcpcb *);
+       tp = malloc(sizeof(*tp), M_PCB, M_NOWAIT);
+       if (tp == NULL)
+               return ((struct tcpcb *)0);
+       bzero((char *) tp, sizeof(struct tcpcb));

        tp->seg_next = tp->seg_prev = (struct tcpiphdr *)tp;
        tp->seg_next = tp->seg_prev = (struct tcpiphdr *)tp;

-       tp->t_maxseg = 1024;
+       tp->t_maxseg = tcp_mssdflt;
+
+       tp->t_flags = tcp_do_rfc1323 ? (TF_REQ_SCALE|TF_REQ_TSTMP) : 0;

        tp->t_inpcb = inp;
        tp->t_inpcb = inp;

+       /*
+        * Init srtt to TCPTV_SRTTBASE (0), so we can tell that we have no
+        * rtt estimate.  Set rttvar so that srtt + 2 * rttvar gives
+        * reasonable initial retransmit time.
+        */
+       tp->t_srtt = TCPTV_SRTTBASE;
+       tp->t_rttvar = tcp_rttdflt * PR_SLOWHZ << 2;
+       tp->t_rttmin = TCPTV_MIN;
+       TCPT_RANGESET(tp->t_rxtcur, 
+           ((TCPTV_SRTTBASE >> 2) + (TCPTV_SRTTDFLT << 2)) >> 1,
+           TCPTV_MIN, TCPTV_REXMTMAX);
+       tp->snd_cwnd = TCP_MAXWIN << TCP_MAX_WINSHIFT;
+       tp->snd_ssthresh = TCP_MAXWIN << TCP_MAX_WINSHIFT;
+       inp->inp_ip.ip_ttl = ip_defttl;

        inp->inp_ppcb = (caddr_t)tp;
        return (tp);
 }
        inp->inp_ppcb = (caddr_t)tp;
        return (tp);
 }


@@ -162,19 +241,23 @@ COUNT(TCP_NEWTCPCB);
  * the specified error.  If connection is synchronized,
  * then send a RST to peer.
  */
  * the specified error.  If connection is synchronized,
  * then send a RST to peer.
  */

+struct tcpcb *

 tcp_drop(tp, errno)
 tcp_drop(tp, errno)

-       struct tcpcb *tp;
+       register struct tcpcb *tp;

        int errno;
 {
        struct socket *so = tp->t_inpcb->inp_socket;
 
        int errno;
 {
        struct socket *so = tp->t_inpcb->inp_socket;
 

-COUNT(TCP_DROP);

        if (TCPS_HAVERCVDSYN(tp->t_state)) {
                tp->t_state = TCPS_CLOSED;
        if (TCPS_HAVERCVDSYN(tp->t_state)) {
                tp->t_state = TCPS_CLOSED;

-               tcp_output(tp);
-       }
+               (void) tcp_output(tp);
+               tcpstat.tcps_drops++;
+       } else
+               tcpstat.tcps_conndrops++;
+       if (errno == ETIMEDOUT && tp->t_softerror)
+               errno = tp->t_softerror;

        so->so_error = errno;
        so->so_error = errno;

-       tcp_close(tp);
+       return (tcp_close(tp));

 }
 
 /*
 }
 
 /*


@@ -183,39 +266,180 @@ COUNT(TCP_DROP);
  *     discard internet protocol block
  *     wake up any sleepers
  */
  *     discard internet protocol block
  *     wake up any sleepers
  */

+struct tcpcb *

 tcp_close(tp)
        register struct tcpcb *tp;
 {
        register struct tcpiphdr *t;
        struct inpcb *inp = tp->t_inpcb;
        struct socket *so = inp->inp_socket;
 tcp_close(tp)
        register struct tcpcb *tp;
 {
        register struct tcpiphdr *t;
        struct inpcb *inp = tp->t_inpcb;
        struct socket *so = inp->inp_socket;

+       register struct mbuf *m;
+#ifdef RTV_RTT
+       register struct rtentry *rt;
+
+       /*
+        * If we sent enough data to get some meaningful characteristics,
+        * save them in the routing entry.  'Enough' is arbitrarily 
+        * defined as the sendpipesize (default 4K) * 16.  This would
+        * give us 16 rtt samples assuming we only get one sample per
+        * window (the usual case on a long haul net).  16 samples is
+        * enough for the srtt filter to converge to within 5% of the correct
+        * value; fewer samples and we could save a very bogus rtt.
+        *
+        * Don't update the default route's characteristics and don't
+        * update anything that the user "locked".
+        */
+       if (SEQ_LT(tp->iss + so->so_snd.sb_hiwat * 16, tp->snd_max) &&
+           (rt = inp->inp_route.ro_rt) &&
+           ((struct sockaddr_in *)rt_key(rt))->sin_addr.s_addr != INADDR_ANY) {
+               register u_long i;

 
 

-COUNT(TCP_CLOSE);
+               if ((rt->rt_rmx.rmx_locks & RTV_RTT) == 0) {
+                       i = tp->t_srtt *
+                           (RTM_RTTUNIT / (PR_SLOWHZ * TCP_RTT_SCALE));
+                       if (rt->rt_rmx.rmx_rtt && i)
+                               /*
+                                * filter this update to half the old & half
+                                * the new values, converting scale.
+                                * See route.h and tcp_var.h for a
+                                * description of the scaling constants.
+                                */
+                               rt->rt_rmx.rmx_rtt =
+                                   (rt->rt_rmx.rmx_rtt + i) / 2;
+                       else
+                               rt->rt_rmx.rmx_rtt = i;
+               }
+               if ((rt->rt_rmx.rmx_locks & RTV_RTTVAR) == 0) {
+                       i = tp->t_rttvar *
+                           (RTM_RTTUNIT / (PR_SLOWHZ * TCP_RTTVAR_SCALE));
+                       if (rt->rt_rmx.rmx_rttvar && i)
+                               rt->rt_rmx.rmx_rttvar =
+                                   (rt->rt_rmx.rmx_rttvar + i) / 2;
+                       else
+                               rt->rt_rmx.rmx_rttvar = i;
+               }
+               /*
+                * update the pipelimit (ssthresh) if it has been updated
+                * already or if a pipesize was specified & the threshhold
+                * got below half the pipesize.  I.e., wait for bad news
+                * before we start updating, then update on both good
+                * and bad news.
+                */
+               if ((rt->rt_rmx.rmx_locks & RTV_SSTHRESH) == 0 &&
+                   (i = tp->snd_ssthresh) && rt->rt_rmx.rmx_ssthresh ||
+                   i < (rt->rt_rmx.rmx_sendpipe / 2)) {
+                       /*
+                        * convert the limit from user data bytes to
+                        * packets then to packet data bytes.
+                        */
+                       i = (i + tp->t_maxseg / 2) / tp->t_maxseg;
+                       if (i < 2)
+                               i = 2;
+                       i *= (u_long)(tp->t_maxseg + sizeof (struct tcpiphdr));
+                       if (rt->rt_rmx.rmx_ssthresh)
+                               rt->rt_rmx.rmx_ssthresh =
+                                   (rt->rt_rmx.rmx_ssthresh + i) / 2;
+                       else
+                               rt->rt_rmx.rmx_ssthresh = i;
+               }
+       }
+#endif /* RTV_RTT */
+       /* free the reassembly queue, if any */

        t = tp->seg_next;
        t = tp->seg_next;

-       for (; t != (struct tcpiphdr *)tp; t = (struct tcpiphdr *)t->ti_next)
-               m_freem(dtom(t));
+       while (t != (struct tcpiphdr *)tp) {
+               t = (struct tcpiphdr *)t->ti_next;
+               m = REASS_MBUF((struct tcpiphdr *)t->ti_prev);
+               remque(t->ti_prev);
+               m_freem(m);
+       }

        if (tp->t_template)
                (void) m_free(dtom(tp->t_template));
        if (tp->t_template)
                (void) m_free(dtom(tp->t_template));

-       if (tp->t_tcpopt)
-               (void) m_free(dtom(tp->t_tcpopt));
-       if (tp->t_ipopt)
-               (void) m_free(dtom(tp->t_ipopt));
-       (void) m_free(dtom(tp));
+       free(tp, M_PCB);

        inp->inp_ppcb = 0;
        inp->inp_ppcb = 0;

-       in_pcbdetach(inp);

        soisdisconnected(so);
        soisdisconnected(so);

+       /* clobber input pcb cache if we're closing the cached connection */
+       if (inp == tcp_last_inpcb)
+               tcp_last_inpcb = &tcb;
+       in_pcbdetach(inp);
+       tcpstat.tcps_closed++;
+       return ((struct tcpcb *)0);

 }
 
 }
 

+void

 tcp_drain()
 {
 
 tcp_drain()
 {
 

-COUNT(TCP_DRAIN);

 }
 
 }
 

-tcp_ctlinput(m)
-       struct mbuf *m;
+/*
+ * Notify a tcp user of an asynchronous error;
+ * store error as soft error, but wake up user
+ * (for now, won't do anything until can select for soft error).
+ */
+void
+tcp_notify(inp, error)
+       struct inpcb *inp;
+       int error;
+{
+       register struct tcpcb *tp = (struct tcpcb *)inp->inp_ppcb;
+       register struct socket *so = inp->inp_socket;
+
+       /*
+        * Ignore some errors if we are hooked up.
+        * If connection hasn't completed, has retransmitted several times,
+        * and receives a second error, give up now.  This is better
+        * than waiting a long time to establish a connection that
+        * can never complete.
+        */
+       if (tp->t_state == TCPS_ESTABLISHED &&
+            (error == EHOSTUNREACH || error == ENETUNREACH ||
+             error == EHOSTDOWN)) {
+               return;
+       } else if (tp->t_state < TCPS_ESTABLISHED && tp->t_rxtshift > 3 &&
+           tp->t_softerror)
+               so->so_error = error;
+       else 
+               tp->t_softerror = error;
+       wakeup((caddr_t) &so->so_timeo);
+       sorwakeup(so);
+       sowwakeup(so);
+}
+
+void
+tcp_ctlinput(cmd, sa, ip)
+       int cmd;
+       struct sockaddr *sa;
+       register struct ip *ip;
+{
+       register struct tcphdr *th;
+       extern struct in_addr zeroin_addr;
+       extern u_char inetctlerrmap[];
+       void (*notify) __P((struct inpcb *, int)) = tcp_notify;
+
+       if (cmd == PRC_QUENCH)
+               notify = tcp_quench;
+       else if (!PRC_IS_REDIRECT(cmd) &&
+                ((unsigned)cmd > PRC_NCMDS || inetctlerrmap[cmd] == 0))
+               return;
+       if (ip) {
+               th = (struct tcphdr *)((caddr_t)ip + (ip->ip_hl << 2));
+               in_pcbnotify(&tcb, sa, th->th_dport, ip->ip_src, th->th_sport,
+                       cmd, notify);
+       } else
+               in_pcbnotify(&tcb, sa, 0, zeroin_addr, 0, cmd, notify);
+}
+
+/*
+ * When a source quench is received, close congestion window
+ * to one segment.  We will gradually open it again as we proceed.
+ */
+void
+tcp_quench(inp, errno)
+       struct inpcb *inp;
+       int errno;

 {
 {

+       struct tcpcb *tp = intotcpcb(inp);

 
 

-COUNT(TCP_CTLINPUT);
-       m_freem(m);
+       if (tp)
+               tp->snd_cwnd = tp->t_maxseg;

 }
 }