Implement 4.4's SUGID flag, which keeps track of processes which

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

/*
 * Copyright (c) 1982, 1986, 1988, 1990 Regents of the University of California.
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. All advertising materials mentioning features or use of this software
 *    must display the following acknowledgement:
 *	This product includes software developed by the University of
 *	California, Berkeley and its contributors.
 * 4. Neither the name of the University nor the names of its contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 *
 *	from: @(#)tcp_subr.c	7.20 (Berkeley) 12/1/90
 *	$Id: tcp_subr.c,v 1.5 1993/11/25 01:35:16 wollman Exp $
 */

#include "param.h"
#include "systm.h"
#include "malloc.h"
#include "mbuf.h"
#include "socket.h"
#include "socketvar.h"
#include "protosw.h"
#include "errno.h"

#include "../net/route.h"
#include "../net/if.h"

#include "in.h"
#include "in_systm.h"
#include "ip.h"
#include "in_pcb.h"
#include "in_var.h"
#include "ip_var.h"
#include "ip_icmp.h"
#include "tcp.h"
#include "tcp_fsm.h"
#include "tcp_seq.h"
#include "tcp_timer.h"
#include "tcp_var.h"
#include "tcpip.h"

tcp_seq	tcp_iss;
struct	inpcb tcb;
struct	tcpstat tcpstat;

#ifdef MTUDISC
int tcp_mtuchanged(struct inpcb *, int);
#endif

/*
 * Tcp initialization
 */
void
tcp_init()
{

	tcp_iss = 1;		/* wrong */
	tcb.inp_next = tcb.inp_prev = &tcb;
	if (max_protohdr < sizeof(struct tcpiphdr))
		max_protohdr = sizeof(struct tcpiphdr);
	if (max_linkhdr + sizeof(struct tcpiphdr) > MHLEN)
		panic("tcp_init");
}

/*
 * Create template to be used to send tcp packets on a connection.
 * Call after host entry created, allocates an mbuf and fills
 * in a skeletal tcp/ip header, minimizing the amount of work
 * necessary when the connection is used.
 */
struct tcpiphdr *
tcp_template(tp)
	struct tcpcb *tp;
{
	register struct inpcb *inp = tp->t_inpcb;
	register struct mbuf *m;
	register struct tcpiphdr *n;

	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_len = htons(sizeof (struct tcpiphdr) - sizeof (struct ip));
	n->ti_src = inp->inp_laddr;
	n->ti_dst = inp->inp_faddr;
	n->ti_sport = inp->inp_lport;
	n->ti_dport = inp->inp_fport;
	n->ti_seq = 0;
	n->ti_ack = 0;
	n->ti_x2 = 0;
	n->ti_off = 5;
	n->ti_flags = 0;
	n->ti_win = 0;
	n->ti_sum = 0;
	n->ti_urp = 0;
	return (n);
}

/*
 * Send a single message to the TCP at address specified by
 * the given TCP/IP header.  If m == 0, then we make a copy
 * of the tcpiphdr at ti and send directly to the addressed host.
 * This is used to force keep alive messages out using the TCP
 * template for a connection tp->t_template.  If flags are given
 * then we send a message back to the TCP which originated the
 * segment ti, and discard the mbuf containing it and any other
 * attached mbufs.
 *
 * In any case the ack and sequence number of the transmitted
 * segment are as specified by the parameters.
 */
void
tcp_respond(tp, ti, m, ack, seq, flags)
	struct tcpcb *tp;
	register struct tcpiphdr *ti;
	register struct mbuf *m;
	tcp_seq ack, seq;
	int flags;
{
	register int tlen;
	int win = 0;
	struct route *ro = 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;
#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 {
		m_freem(m->m_next);
		m->m_next = 0;
		m->m_data = (caddr_t)ti;
		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
	}
	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_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_win = htons((u_short)win);
	ti->ti_urp = 0;
	ti->ti_sum = in_cksum(m, tlen);
	((struct ip *)ti)->ip_len = tlen;
	((struct ip *)ti)->ip_ttl = tcp_ttl;
	(void) ip_output(m, (struct mbuf *)0, ro, 0);
}

/*
 * Create a new TCP control block, making an
 * empty reassembly queue and hooking it to the argument
 * protocol control block.
 */
struct tcpcb *
tcp_newtcpcb(inp)
	struct inpcb *inp;
{
	struct mbuf *m = m_getclr(M_DONTWAIT, MT_PCB);
	register struct tcpcb *tp;

	if (m == NULL)
		return ((struct tcpcb *)0);
	tp = mtod(m, struct tcpcb *);
	tp->seg_next = tp->seg_prev = (struct tcpiphdr *)tp;
	tp->t_maxseg = tcp_mssdflt;

	tp->t_flags = 0;		/* sends options! */
	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;
	tp->snd_ssthresh = TCP_MAXWIN;
	inp->inp_ip.ip_ttl = tcp_ttl;
	inp->inp_ppcb = (caddr_t)tp;
#ifdef MTUDISC
	/*
	 * Enable Path MTU Discovery on this PCB.
	 */
	inp->inp_mtunotify = tcp_mtuchanged;
	inp->inp_flags |= INP_DISCOVERMTU;
#endif /* MTUDISC */
	return (tp);
}

/*
 * Drop a TCP connection, reporting
 * the specified error.  If connection is synchronized,
 * then send a RST to peer.
 */
struct tcpcb *
tcp_drop(tp, errno)
	register struct tcpcb *tp;
	int errno;
{
	struct socket *so = tp->t_inpcb->inp_socket;

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

/*
 * Close a TCP control block:
 *	discard all space held by the tcp
 *	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;
	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 = 0;

		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;
	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));
	(void) m_free(dtom(tp));
	inp->inp_ppcb = 0;
	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()
{

}

/*
 * 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)
	register struct inpcb *inp;
	int error;
{

	((struct tcpcb *)inp->inp_ppcb)->t_softerror = error;
	wakeup((caddr_t) &inp->inp_socket->so_timeo);
	sorwakeup(inp->inp_socket);
	sowwakeup(inp->inp_socket);
}

/*
 * When we get a PRC_MSGSIZE error (generated by the ICMP layer upon
 * receipt of an ICMP_UNREACH_NEEDFRAG message), we need to get the
 * IP layer to check the cached MTU data that it has in its PCBs.
 * If things have changed, this will cause us to receive a
 * PRC_MTUCHANGED message for /every/ connection to the same
 * destination; that is handled by he tcp_mtuchanged() function,
 * below.
 *
 * In the immortal words of Ken and Dennis, ``You are not expected to
 * understand this.''
 */
void
tcp_checkmtu(struct inpcb *inp, int error) {
#ifdef MTUDISC
  /*
   * XXX - this should also cause an immediate retransmission and
   * slow start, since we know for a fact that the message we just sent
   * got dropped on the floor.  For now, just do what tcp_quench does.
   */
  tcp_quench(inp);
  in_pcbmtu(inp);
#endif /* MTUDISC */
}

#ifdef MTUDISC
int /* grrr... should be void... */
tcp_mtuchanged(struct inpcb *inp, int error) {
  /* don't do anything just yet... */;
}
#endif /* MTUDISC */

void
tcp_ctlinput(cmd, sa, ip)
	int cmd;
	struct sockaddr *sa;
	register struct ip *ip;
{
	register struct tcphdr *th;
	void (*notify)(struct inpcb *, int) = tcp_notify;

	if (cmd == PRC_QUENCH)
		notify = tcp_quench;
#ifdef MTUDISC
	else if (cmd == PRC_MSGSIZE)
		notify = tcp_checkmtu;
	else if (cmd == PRC_MTUCHANGED)	/* just in case */
		notify = tcp_mtuchanged;
#endif /* MTUDISC */
	else if ((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);

	if (tp)
		tp->snd_cwnd = tp->t_maxseg;
}