tcp_usrreq.c

ReactOS是一些高手根据Windows XP的内核编写出的类XP。内核实现机理和API函数调用几乎相同。甚至可以兼容XP的程序。喜欢研究系统内核的人可以看一看。

	default:
		panic("tcp_usrreq");
	}
#ifdef TCPDEBUG
	if (tp && (so->so_options & SO_DEBUG))
		tcp_trace(TA_USER, ostate, tp, (struct tcpiphdr *)0, req);
#endif
	splx(s);
	return (error);
}

/*
 * Common subroutine to open a TCP connection to remote host specified
 * by struct sockaddr_in in mbuf *nam.  Call in_pcbbind to assign a local
 * port number if needed.  Call in_pcbladdr to do the routing and to choose
 * a local host address (interface).  If there is an existing incarnation
 * of the same connection in TIME-WAIT state and if the remote host was
 * sending CC options and if the connection duration was < MSL, then
 * truncate the previous TIME-WAIT state and proceed.
 * Initialize connection parameters and enter SYN-SENT state.
 */
int
tcp_connect(tp, nam)
	register struct tcpcb *tp;
	struct mbuf *nam;
{
	struct inpcb *inp = tp->t_inpcb, *oinp;
	struct socket *so = inp->inp_socket;
	struct tcpcb *otp;
	struct sockaddr_in *sin = mtod(nam, struct sockaddr_in *);
	struct sockaddr_in *ifaddr;
	int error;
	struct rmxp_tao *taop;
	struct rmxp_tao tao_noncached;

	OS_DbgPrint(OSK_MID_TRACE,("Called\n"));

	if (inp->inp_lport == 0) {
		error = in_pcbbind(inp, NULL);
		if (error)
			return error;
	}

	/*
	 * Cannot simply call in_pcbconnect, because there might be an
	 * earlier incarnation of this same connection still in
	 * TIME_WAIT state, creating an ADDRINUSE error.
	 */
	error = in_pcbladdr(inp, nam, &ifaddr);
        if (error) {
	    OS_DbgPrint(OSK_MID_TRACE,("leaving %d\n", error));
	    return error;
	}
	oinp = in_pcblookup(inp->inp_pcbinfo->listhead,
	    sin->sin_addr, sin->sin_port,
	    inp->inp_laddr.s_addr != INADDR_ANY ? inp->inp_laddr
						: ifaddr->sin_addr,
	    inp->inp_lport,  0);
	if (oinp) {
		if (oinp != inp && (otp = intotcpcb(oinp)) != NULL &&
		otp->t_state == TCPS_TIME_WAIT &&
		    otp->t_duration < TCPTV_MSL &&
		    (otp->t_flags & TF_RCVD_CC))
			otp = tcp_close(otp);
		else {
		    OS_DbgPrint(OSK_MID_TRACE,("leaving EADDRINUSE\n"));
		    return EADDRINUSE;
		}
	}
	if (inp->inp_laddr.s_addr == INADDR_ANY)
		inp->inp_laddr = ifaddr->sin_addr;
	inp->inp_faddr = sin->sin_addr;
	inp->inp_fport = sin->sin_port;
	in_pcbrehash(inp);

	tp->t_template = tcp_template(tp);
	if (tp->t_template == 0) {
		in_pcbdisconnect(inp);
		OS_DbgPrint(OSK_MID_TRACE,("Leaving ENOBUFS\n"));
		return ENOBUFS;
	}

	/* Compute window scaling to request.  */
	while (tp->request_r_scale < TCP_MAX_WINSHIFT &&
	    (TCP_MAXWIN << tp->request_r_scale) < so->so_rcv.sb_hiwat)
		tp->request_r_scale++;

	soisconnecting(so);
	tcpstat.tcps_connattempt++;
	tp->t_state = TCPS_SYN_SENT;
	tp->t_timer[TCPT_KEEP] = tcp_keepinit;
	tp->iss = tcp_iss; tcp_iss += TCP_ISSINCR/2;
	tcp_sendseqinit(tp);

	/*
	 * Generate a CC value for this connection and
	 * check whether CC or CCnew should be used.
	 */
	if ((taop = tcp_gettaocache(tp->t_inpcb)) == NULL) {
		taop = &tao_noncached;
		bzero(taop, sizeof(*taop));
	}

	tp->cc_send = CC_INC(tcp_ccgen);
	if (taop->tao_ccsent != 0 &&
	    CC_GEQ(tp->cc_send, taop->tao_ccsent)) {
		taop->tao_ccsent = tp->cc_send;
	} else {
		taop->tao_ccsent = 0;
		tp->t_flags |= TF_SENDCCNEW;
	}

	OS_DbgPrint(OSK_MID_TRACE,("Leaving 0\n"));
	return 0;
}

int
tcp_ctloutput(op, so, level, optname, mp)
	int op;
	struct socket *so;
	int level, optname;
	struct mbuf **mp;
{
	int error = 0, s;
	struct inpcb *inp;
	register struct tcpcb *tp;
	register struct mbuf *m;
	register int i;

	s = splnet();
	inp = sotoinpcb(so);
	if (inp == NULL) {
		splx(s);
		if (op == PRCO_SETOPT && *mp)
			(void) m_free(*mp);
		return (ECONNRESET);
	}
	if (level != IPPROTO_TCP) {
		error = ip_ctloutput(op, so, level, optname, mp);
		splx(s);
		return (error);
	}
	tp = intotcpcb(inp);

	switch (op) {

	case PRCO_SETOPT:
		m = *mp;
		switch (optname) {

		case TCP_NODELAY:
			if (m == NULL || m->m_len < sizeof (int))
				error = EINVAL;
			else if (*mtod(m, int *))
				tp->t_flags |= TF_NODELAY;
			else
				tp->t_flags &= ~TF_NODELAY;
			break;

		case TCP_MAXSEG:
			if (m && (i = *mtod(m, int *)) > 0 && i <= tp->t_maxseg)
				tp->t_maxseg = i;
			else
				error = EINVAL;
			break;

		case TCP_NOOPT:
			if (m == NULL || m->m_len < sizeof (int))
				error = EINVAL;
			else if (*mtod(m, int *))
				tp->t_flags |= TF_NOOPT;
			else
				tp->t_flags &= ~TF_NOOPT;
			break;

		case TCP_NOPUSH:
			if (m == NULL || m->m_len < sizeof (int))
				error = EINVAL;
			else if (*mtod(m, int *))
				tp->t_flags |= TF_NOPUSH;
			else
				tp->t_flags &= ~TF_NOPUSH;
			break;

		default:
			error = ENOPROTOOPT;
			break;
		}
		if (m)
			(void) m_free(m);
		break;

	case PRCO_GETOPT:
		*mp = m = m_get(M_WAIT, MT_SOOPTS);
		m->m_len = sizeof(int);

		switch (optname) {
		case TCP_NODELAY:
			*mtod(m, int *) = tp->t_flags & TF_NODELAY;
			break;
		case TCP_MAXSEG:
			*mtod(m, int *) = tp->t_maxseg;
			break;
		case TCP_NOOPT:
			*mtod(m, int *) = tp->t_flags & TF_NOOPT;
			break;
		case TCP_NOPUSH:
			*mtod(m, int *) = tp->t_flags & TF_NOPUSH;
			break;
		default:
			error = ENOPROTOOPT;
			break;
		}
		break;
	}
	splx(s);
	return (error);
}

/*
 * tcp_sendspace and tcp_recvspace are the default send and receive window
 * sizes, respectively.  These are obsolescent (this information should
 * be set by the route).
 */
u_long	tcp_sendspace = 1024*16;
u_long	tcp_recvspace = 1024*16;

/*
 * Attach TCP protocol to socket, allocating
 * internet protocol control block, tcp control block,
 * bufer space, and entering LISTEN state if to accept connections.
 */
int
tcp_attach(so)
	struct socket *so;
{
	register struct tcpcb *tp;
	struct inpcb *inp;
	int error;

	if (so->so_snd.sb_hiwat == 0 || so->so_rcv.sb_hiwat == 0) {
		error = soreserve(so, tcp_sendspace, tcp_recvspace);
		if (error)
			return (error);
	}
	error = in_pcballoc(so, &tcbinfo);
	if (error)
		return (error);
	inp = sotoinpcb(so);
	tp = tcp_newtcpcb(inp);
	if (tp == 0) {
		int nofd = so->so_state & SS_NOFDREF;	/* XXX */

		so->so_state &= ~SS_NOFDREF;	/* don't free the socket yet */
		in_pcbdetach(inp);
		so->so_state |= nofd;
		return (ENOBUFS);
	}
	tp->t_state = TCPS_CLOSED;
	return (0);
}

/*
 * Initiate (or continue) disconnect.
 * If embryonic state, just send reset (once).
 * If in ``let data drain'' option and linger null, just drop.
 * Otherwise (hard), mark socket disconnecting and drop
 * current input data; switch states based on user close, and
 * send segment to peer (with FIN).
 */
struct tcpcb *
tcp_disconnect(tp)
	register struct tcpcb *tp;
{
	struct socket *so = tp->t_inpcb->inp_socket;

	if (tp->t_state < TCPS_ESTABLISHED)
		tp = tcp_close(tp);
	else if ((so->so_options & SO_LINGER) && so->so_linger == 0)
		tp = tcp_drop(tp, 0);
	else {
		soisdisconnecting(so);
		sbflush(&so->so_rcv);
		tp = tcp_usrclosed(tp);
		if (tp)
			(void) tcp_output(tp);
	}
	return (tp);
}

/*
 * User issued close, and wish to trail through shutdown states:
 * if never received SYN, just forget it.  If got a SYN from peer,
 * but haven't sent FIN, then go to FIN_WAIT_1 state to send peer a FIN.
 * If already got a FIN from peer, then almost done; go to LAST_ACK
 * state.  In all other cases, have already sent FIN to peer (e.g.
 * after PRU_SHUTDOWN), and just have to play tedious game waiting
 * for peer to send FIN or not respond to keep-alives, etc.
 * We can let the user exit from the close as soon as the FIN is acked.
 */
struct tcpcb *
tcp_usrclosed(tp)
	register struct tcpcb *tp;
{

	switch (tp->t_state) {

	case TCPS_CLOSED:
	case TCPS_LISTEN:
		tp->t_state = TCPS_CLOSED;
		tp = tcp_close(tp);
		break;

	case TCPS_SYN_SENT:
	case TCPS_SYN_RECEIVED:
		tp->t_flags |= TF_NEEDFIN;
		break;

	case TCPS_ESTABLISHED:
		tp->t_state = TCPS_FIN_WAIT_1;
		break;

	case TCPS_CLOSE_WAIT:
		tp->t_state = TCPS_LAST_ACK;
		break;
	}
	if (tp && tp->t_state >= TCPS_FIN_WAIT_2) {
		soisdisconnected(tp->t_inpcb->inp_socket);
		/* To prevent the connection hanging in FIN_WAIT_2 forever. */
		if (tp->t_state == TCPS_FIN_WAIT_2)
			tp->t_timer[TCPT_2MSL] = tcp_maxidle;
	}
	return (tp);
}

/*
 * Sysctl for tcp variables.
 */
int
tcp_sysctl(name, namelen, oldp, oldlenp, newp, newlen)
	int *name;
	u_int namelen;
	void *oldp;
	size_t *oldlenp;
	void *newp;
	size_t newlen;
{
	/* All sysctl names at this level are terminal. */
	if (namelen != 1)
		return (ENOTDIR);

	switch (name[0]) {
	case TCPCTL_DO_RFC1323:
		return (sysctl_int(oldp, oldlenp, newp, newlen,
		    &tcp_do_rfc1323));
	case TCPCTL_DO_RFC1644:
		return (sysctl_int(oldp, oldlenp, newp, newlen,
		    &tcp_do_rfc1644));
	case TCPCTL_MSSDFLT:
		return (sysctl_int(oldp, oldlenp, newp, newlen,
		    &tcp_mssdflt));
	case TCPCTL_STATS:
		return (sysctl_rdstruct(oldp, oldlenp, newp, &tcpstat,
					sizeof tcpstat));
	case TCPCTL_RTTDFLT:
		return (sysctl_int(oldp, oldlenp, newp, newlen, &tcp_rttdflt));
	case TCPCTL_KEEPIDLE:
		return (sysctl_int(oldp, oldlenp, newp, newlen,
				   &tcp_keepidle));
	case TCPCTL_KEEPINTVL:
		return (sysctl_int(oldp, oldlenp, newp, newlen,
				   &tcp_keepintvl));
	case TCPCTL_SENDSPACE:
		return (sysctl_int(oldp, oldlenp, newp, newlen,
				   (int *)&tcp_sendspace)); /* XXX */
	case TCPCTL_RECVSPACE:
		return (sysctl_int(oldp, oldlenp, newp, newlen,
				   (int *)&tcp_recvspace)); /* XXX */
	case TCPCTL_KEEPINIT:
		return (sysctl_int(oldp, oldlenp, newp, newlen,
				   &tcp_keepinit));
	default:
		return (ENOPROTOOPT);
	}
	/* NOTREACHED */
}