tcp_subr.c

RTEMS (Real-Time Executive for Multiprocessor Systems) is a free open source real-time operating sys

			 * 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;
			tcpstat.tcps_cachedssthresh++;
		}
	}
	/* free the reassembly queue, if any */
	t = tp->seg_next;
	while (t != (struct tcpiphdr *)tp) {
		t = (struct tcpiphdr *)t->ti_next;
#if (defined(__GNUC__) && defined(__arm__))
        LD32_UNALGN((struct tcpiphdr *)t->ti_prev,m);
#else
		m = REASS_MBUF((struct tcpiphdr *)t->ti_prev);
#endif
		remque(t->ti_prev);
		m_freem(m);
	}
	if (tp->t_template)
		(void) m_free(dtom(tp->t_template));
	free(tp, M_PCB);
	inp->inp_ppcb = 0;
	soisdisconnected(so);
	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).
 */
static 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;
	soconnwakeup (so);
	sorwakeup(so);
	sowwakeup(so);
}

#ifdef __rtems__
#define INP_INFO_RLOCK(a)
#define INP_INFO_RUNLOCK(a)
#define INP_LOCK(a)
#define INP_UNLOCK(a)
#endif

static int
tcp_pcblist(SYSCTL_HANDLER_ARGS)
{
	int error, i, n, s;
	struct inpcb *inp, **inp_list;
	inp_gen_t gencnt;
	struct xinpgen xig;

	/*
	 * The process of preparing the TCB list is too time-consuming and
	 * resource-intensive to repeat twice on every request.
	 */
	if (req->oldptr == 0) {
		n = tcbinfo.ipi_count;
		req->oldidx = 2 * (sizeof xig)
			+ (n + n/8) * sizeof(struct xtcpcb);
		return 0;
	}

	if (req->newptr != 0)
		return EPERM;

	/*
	 * OK, now we're committed to doing something.
	 */
	s = splnet();
	INP_INFO_RLOCK(&tcbinfo);
	gencnt = tcbinfo.ipi_gencnt;
	n = tcbinfo.ipi_count;
	INP_INFO_RUNLOCK(&tcbinfo);
	splx(s);

	sysctl_wire_old_buffer(req, 2 * (sizeof xig)
		+ n * sizeof(struct xtcpcb));

	xig.xig_len = sizeof xig;
	xig.xig_count = n;
	xig.xig_gen = gencnt;
/*	xig.xig_sogen = so_gencnt; remove by ccj */
	error = SYSCTL_OUT(req, &xig, sizeof xig);
	if (error)
		return error;

	/* ccj add exit if the count is 0 */
	if (!n)
		return error;
  
	inp_list = malloc(n * sizeof *inp_list, M_TEMP, M_WAITOK);
	if (inp_list == 0)
		return ENOMEM;
	
	s = splnet();
	INP_INFO_RLOCK(&tcbinfo);
	for (inp = LIST_FIRST(tcbinfo.listhead), i = 0; inp && i < n;
	     inp = LIST_NEXT(inp, inp_list)) {
		INP_LOCK(inp);
		if (inp->inp_gencnt <= gencnt)
#if 0
      &&
		    cr_canseesocket(req->td->td_ucred, inp->inp_socket) == 0)
#endif
			inp_list[i++] = inp;
		INP_UNLOCK(inp);
	}
	INP_INFO_RUNLOCK(&tcbinfo);
	splx(s);
	n = i;

	error = 0;
	for (i = 0; i < n; i++) {
		inp = inp_list[i];
		INP_LOCK(inp);
		if (inp->inp_gencnt <= gencnt) {
			struct xtcpcb xt;
			caddr_t inp_ppcb;
			xt.xt_len = sizeof xt;
			/* XXX should avoid extra copy */
			bcopy(inp, &xt.xt_inp, sizeof *inp);
			inp_ppcb = inp->inp_ppcb;
			if (inp_ppcb != NULL)
				bcopy(inp_ppcb, &xt.xt_tp, sizeof xt.xt_tp);
			else
				bzero((char *) &xt.xt_tp, sizeof xt.xt_tp);
#if 0      
			if (inp->inp_socket)
				sotoxsocket(inp->inp_socket, &xt.xt_socket);
#endif
			error = SYSCTL_OUT(req, &xt, sizeof xt);
		}
		INP_UNLOCK(inp);
	}
	if (!error) {
		/*
		 * Give the user an updated idea of our state.
		 * If the generation differs from what we told
		 * her before, she knows that something happened
		 * while we were processing this request, and it
		 * might be necessary to retry.
		 */
		s = splnet();
		INP_INFO_RLOCK(&tcbinfo);
		xig.xig_gen = tcbinfo.ipi_gencnt;
#if 0    
		xig.xig_sogen = so_gencnt;
#endif
		xig.xig_count = tcbinfo.ipi_count;
		INP_INFO_RUNLOCK(&tcbinfo);
		splx(s);
		error = SYSCTL_OUT(req, &xig, sizeof xig);
	}
	free(inp_list, M_TEMP);
	return error;
}

SYSCTL_PROC(_net_inet_tcp, TCPCTL_PCBLIST, pcblist, CTLFLAG_RD, 0, 0,
	    tcp_pcblist, "S,xtcpcb", "List of active TCP connections");

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

	if (cmd == PRC_QUENCH)
		notify = tcp_quench;
#if 1
	else if (cmd == PRC_MSGSIZE)
		notify = tcp_mtudisc;
#endif
	else if (!PRC_IS_REDIRECT(cmd) &&
		 ((unsigned)cmd > PRC_NCMDS || inetctlerrmap[cmd] == 0))
		return;
	if (ip) {
		th = (struct tcphdr *)((caddr_t)ip 
				       + (IP_VHL_HL(ip->ip_vhl) << 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, errnum)
	struct inpcb *inp;
	int errnum;
{
	struct tcpcb *tp = intotcpcb(inp);

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

#if 1
/*
 * When `need fragmentation' ICMP is received, update our idea of the MSS
 * based on the new value in the route.  Also nudge TCP to send something,
 * since we know the packet we just sent was dropped.
 * This duplicates some code in the tcp_mss() function in tcp_input.c.
 */
void
tcp_mtudisc(inp, errnum)
	struct inpcb *inp;
	int errnum;
{
	struct tcpcb *tp = intotcpcb(inp);
	struct rtentry *rt;
	struct rmxp_tao *taop;
	struct socket *so = inp->inp_socket;
	int offered;
	int mss;

	if (tp) {
		rt = tcp_rtlookup(inp);
		if (!rt || !rt->rt_rmx.rmx_mtu) {
			tp->t_maxopd = tp->t_maxseg = tcp_mssdflt;
			return;
		}
		taop = rmx_taop(rt->rt_rmx);
		offered = taop->tao_mssopt;
		mss = rt->rt_rmx.rmx_mtu - sizeof(struct tcpiphdr);
		if (offered)
			mss = min(mss, offered);
		/*
		 * XXX - The above conditional probably violates the TCP
		 * spec.  The problem is that, since we don't know the
		 * other end's MSS, we are supposed to use a conservative
		 * default.  But, if we do that, then MTU discovery will
		 * never actually take place, because the conservative
		 * default is much less than the MTUs typically seen
		 * on the Internet today.  For the moment, we'll sweep
		 * this under the carpet.
		 *
		 * The conservative default might not actually be a problem
		 * if the only case this occurs is when sending an initial
		 * SYN with options and data to a host we've never talked
		 * to before.  Then, they will reply with an MSS value which
		 * will get recorded and the new parameters should get
		 * recomputed.  For Further Study.
		 */
		if (tp->t_maxopd <= mss)
			return;
		tp->t_maxopd = mss;

		if ((tp->t_flags & (TF_REQ_TSTMP|TF_NOOPT)) == TF_REQ_TSTMP &&
		    (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 &&
		    (tp->t_flags & TF_RCVD_CC) == TF_RCVD_CC)
			mss -= TCPOLEN_CC_APPA;
#if	(MCLBYTES & (MCLBYTES - 1)) == 0
		if (mss > MCLBYTES)
			mss &= ~(MCLBYTES-1);
#else
		if (mss > MCLBYTES)
			mss = mss / MCLBYTES * MCLBYTES;
#endif
		if (so->so_snd.sb_hiwat < mss)
			mss = so->so_snd.sb_hiwat;

		tp->t_maxseg = mss;

		tcpstat.tcps_mturesent++;
		tp->t_rtt = 0;
		tp->snd_nxt = tp->snd_una;
		tcp_output(tp);
	}
}
#endif

/*
 * Look-up the routing entry to the peer of this inpcb.  If no route
 * is found and it cannot be allocated the return NULL.  This routine
 * is called by TCP routines that access the rmx structure and by tcp_mss
 * to get the interface MTU.
 */
struct rtentry *
tcp_rtlookup(inp)
	struct inpcb *inp;
{
	struct route *ro;
	struct rtentry *rt;

	ro = &inp->inp_route;
	rt = ro->ro_rt;
	if (rt == NULL || !(rt->rt_flags & RTF_UP)) {
		/* No route yet, so try to acquire one */
		if (inp->inp_faddr.s_addr != INADDR_ANY) {
			ro->ro_dst.sa_family = AF_INET;
			ro->ro_dst.sa_len = sizeof(ro->ro_dst);
			((struct sockaddr_in *) &ro->ro_dst)->sin_addr =
				inp->inp_faddr;
			rtalloc(ro);
			rt = ro->ro_rt;
		}
	}
	return rt;
}

/*
 * Return a pointer to the cached information about the remote host.
 * The cached information is stored in the protocol specific part of
 * the route metrics.
 */
struct rmxp_tao *
tcp_gettaocache(inp)
	struct inpcb *inp;
{
	struct rtentry *rt = tcp_rtlookup(inp);

	/* Make sure this is a host route and is up. */
	if (rt == NULL ||
	    (rt->rt_flags & (RTF_UP|RTF_HOST)) != (RTF_UP|RTF_HOST))
		return NULL;

	return rmx_taop(rt->rt_rmx);
}

/*
 * Clear all the TAO cache entries, called from tcp_init.
 *
 * XXX
 * This routine is just an empty one, because we assume that the routing
 * routing tables are initialized at the same time when TCP, so there is
 * nothing in the cache left over.
 */
static void
tcp_cleartaocache(void)
{ }