ftp.nice.ch/pub/next/connectivity/protocol/GateKeeper.3.0.Beta.4.s.tar.gz#/GateKeeper.3.0.Beta.4.s/ppp/pppd/sys-NeXT.c

This is sys-NeXT.c in view mode; [Download] [Up]

/*
 * sys-next.c - System-dependent procedures for setting up
 * PPP interfaces on NeXT 3.2/3.3  systems
 *
 * Copyright (c) 1989 Carnegie Mellon University.
 * Copyright (c) 1994 Philippe-Andre Prindeville.
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms are permitted
 * provided that the above copyright notice and this paragraph are
 * duplicated in all such forms and that any documentation,
 * advertising materials, and other materials related to such
 * distribution and use acknowledge that the software was developed
 * by Carnegie Mellon University.  The name of the
 * University may not be used to endorse or promote products derived
 * from this software without specific prior written permission.
 * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR
 * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED
 * WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE.
 */

#ifndef lint
static char rcsid[] = "$Id: sys-NeXT.c,v 1.8 1997/04/30 05:57:14 paulus Exp $";
#endif

#include <stdio.h>
#include <syslog.h>
#include <termios.h>
#include <utmp.h>
#include <unistd.h>
#include <stdlib.h>
#include <strings.h>
#include <sys/types.h>
#include <sys/file.h>
#include <sys/socket.h>
#include <sys/ioctl.h>
#include <sys/time.h>
#include <sys/errno.h>
#include <sys/stat.h>
#include <sys/fcntl.h>

#include <net/if.h>
#include <net/ppp_defs.h>
#include <net/if_ppp.h>
#include <netdb.h>
#include <netinet/in.h>
#include <netinet/in_systm.h>
#include <netinet/in_var.h>
#include <netinet/ip.h>
#include <netinet/if_ether.h>
#include <net/route.h>
#include <netinet/in.h>

#include <netinfo/ni.h>

#include "pppd.h"

static int initdisc = -1;	/* Initial TTY discipline for ppp_fd */
static int initfdflags = -1;	/* Initial file descriptor flags for ppp_fd */
static int ppp_fd = -1;		/* fd which is set to PPP discipline */

extern int errno;

static int	restore_term;	/* 1 => we've munged the terminal */
static struct termios inittermios; /* Initial TTY termios */

static char *lock_file;		/* name of lock file created */

static int loop_slave = -1;
static int loop_master;
static char loop_name[20];

static unsigned char inbuf[512]; /* buffer for chars read from loopback */

static int sockfd;		/* socket for doing interface ioctls */

#if defined(i386) && defined(HAS_BROKEN_IOCTL)
#define	ioctl	myioctl
#endif

static int if_is_up;		/* the interface is currently up */
static u_int32_t ifaddrs[2];	/* local and remote addresses we set */
static u_int32_t default_route_gateway;	/* gateway addr for default route */
static u_int32_t proxy_arp_addr;	/* remote addr for proxy arp */

/* Prototypes for procedures local to this file. */
static int translate_speed __P((int));
static int baud_rate_of __P((int));
static int dodefaultroute __P((u_int32_t, int));
static int get_ether_addr __P((u_int32_t, struct sockaddr *));
static int ether_by_host __P((char *, struct ether_addr *));

							// open a pty pair for loopback
static int openpty(int *master, int *slave, char *loop_name);

#import "../../MiG/Library/pppd_rpc_defs.h"			// pppd RPC related defs
port_t gk_report __P((int));
void gk_report_oneway __P((int));



/*
 * sys_init - System-dependent initialization.
 */
void
sys_init()
{
    openlog("pppd", LOG_PID | LOG_NDELAY, LOG_PPP);
    setlogmask(LOG_UPTO(LOG_INFO));

    /* Get an internet socket for doing socket ioctl's on. */
    if ((sockfd = socket(AF_INET, SOCK_DGRAM, 0)) < 0) {
	syslog(LOG_ERR, "Couldn't create IP socket: %m");
	die(1);
    }
}

/*
 * sys_cleanup - restore any system state we modified before exiting:
 * mark the interface down, delete default route and/or proxy arp entry.
 * This should call die() because it's called from die().
 */
void
sys_cleanup()
{
    struct ifreq ifr;

    if (if_is_up) {
	strncpy(ifr.ifr_name, ifname, sizeof(ifr.ifr_name));
	if (ioctl(sockfd, SIOCGIFFLAGS, &ifr) >= 0
	    && ((ifr.ifr_flags & IFF_UP) != 0)) {
	    ifr.ifr_flags &= ~IFF_UP;
	    ioctl(sockfd, SIOCSIFFLAGS, &ifr);
	}
    }
//    if (ifaddrs[0] != 0)					// enabled in bsd
//	cifaddr(0, ifaddrs[0], ifaddrs[1]);
    if (default_route_gateway)
	cifdefaultroute(0, 0, default_route_gateway);
    if (proxy_arp_addr)
	cifproxyarp(0, proxy_arp_addr);
}

/*
 * sys_close - Clean up in a child process before execing.
 */
void
sys_close()
{
    close(sockfd);
    if (loop_slave >= 0) {
	close(loop_slave);
	close(loop_master);
    }
    closelog();
}

/*
 * sys_check_options - check the options that the user specified
 */
void
sys_check_options()
{
}

/*
 * ppp_available - check whether the system has any ppp interfaces
 * (in fact we check whether we can do an ioctl on ppp0).
 */
int
ppp_available()
{
    int s, ok;
    struct ifreq ifr;
    extern char *no_ppp_msg;

    if ((s = socket(AF_INET, SOCK_DGRAM, 0)) < 0)
	return 1;		/* can't tell */

    strncpy(ifr.ifr_name, "ppp0", sizeof (ifr.ifr_name));
    ok = ioctl(s, SIOCGIFFLAGS, (caddr_t) &ifr) >= 0;
    close(s);

    no_ppp_msg = "PPP LKS not loaded.\n";

    return ok;
}

/*
 * establish_ppp - Turn the serial port into a ppp interface.
 */
void
establish_ppp(fd)
    int fd;
{
    int pppdisc = PPPDISC;
    int x;

    if (demand) {
	/*
	 * Demand mode - prime the old ppp device to relinquish the unit.
	 */
	if (ioctl(ppp_fd, PPPIOCXFERUNIT, 0) < 0) {
	    syslog(LOG_ERR, "ioctl(transfer ppp unit): %m");
//	    die(1);
	}
    }

    /*
     * Save the old line discipline of fd, and set it to PPP.
     */
    if (ioctl(fd, TIOCGETD, &initdisc) < 0) {
	syslog(LOG_ERR, "ioctl(TIOCGETD): %m");
	die(1);
    }
    if (ioctl(fd, TIOCSETD, &pppdisc) < 0) {
	syslog(LOG_ERR, "ioctl(establish TIOCSETD): %m");
	die(1);
    }

    if (!demand) {
	/*
	 * Find out which interface we were given.
	 */
	if (ioctl(fd, PPPIOCGUNIT, &ifunit) < 0) {	
	    syslog(LOG_ERR, "ioctl(PPPIOCGUNIT): %m");
	    die(1);
	}
    } else {
	/*
	 * Check that we got the same unit again.
	 */
	if (ioctl(fd, PPPIOCGUNIT, &x) < 0) {	
	    syslog(LOG_ERR, "ioctl(PPPIOCGUNIT): %m");
	    die(1);
	}
	if (x != ifunit) {
	    syslog(LOG_ERR, "transfer_ppp failed: wanted unit %d, got %d",
		   ifunit, x);
	    die(1);
	}
	x = NTTYDISC;
	ioctl(loop_slave, TIOCSETD, &x);
    }

    ppp_fd = fd;

    /*
     * Enable debug in the driver if requested.
     */
    if (kdebugflag) {
	if (ioctl(fd, PPPIOCGFLAGS, (caddr_t) &x) < 0) {
	    syslog(LOG_WARNING, "ioctl(PPPIOCGFLAGS): %m");
	} else {
	    x |= (kdebugflag & 0xFF) * SC_DEBUG;
	    if (ioctl(fd, PPPIOCSFLAGS, (caddr_t) &x) < 0)
		syslog(LOG_WARNING, "ioctl(PPPIOCSFLAGS): %m");
	}
    }

    /*
     * Set device for non-blocking reads so PPPD can poll for
     * input from the kernel.
     */
    if ((initfdflags = fcntl(fd, F_GETFL)) == -1
	|| fcntl(fd, F_SETFL, initfdflags | O_NONBLOCK) == -1) {
	syslog(LOG_WARNING, "Couldn't set device to non-blocking mode: %m");
    }
}

/*
 * disestablish_ppp - Restore the serial port to normal operation.
 * This shouldn't call die() because it's called from die().
 */
void
disestablish_ppp(fd)
    int fd;
{
    /* Reset non-blocking mode on fd. */
    if (initfdflags != -1 && fcntl(fd, F_SETFL, initfdflags) < 0)
	syslog(LOG_WARNING, "Couldn't restore device fd flags: %m");
    initfdflags = -1;

    /* Restore old line discipline. */
    if (initdisc >= 0 && ioctl(fd, TIOCSETD, &initdisc) < 0)
	syslog(LOG_ERR, "ioctl(TIOCSETD): %m");
    initdisc = -1;

    if (fd == ppp_fd)
	ppp_fd = -1;
}

/*
 * Check whether the link seems not to be 8-bit clean.
 */
void
clean_check()
{
    int x;
    char *s;

    if (ioctl(ppp_fd, PPPIOCGFLAGS, (caddr_t) &x) == 0) {
	s = NULL;
	switch (~x & (SC_RCV_B7_0|SC_RCV_B7_1|SC_RCV_EVNP|SC_RCV_ODDP)) {
	case SC_RCV_B7_0:
	    s = "bit 7 set to 1";
	    break;
	case SC_RCV_B7_1:
	    s = "bit 7 set to 0";
	    break;
	case SC_RCV_EVNP:
	    s = "odd parity";
	    break;
	case SC_RCV_ODDP:
	    s = "even parity";
	    break;
	}
	if (s != NULL) {
	    syslog(LOG_WARNING, "Serial link is not 8-bit clean:");
	    syslog(LOG_WARNING, "All received characters had %s", s);
	}
    }
}

/*
 * set_up_tty: Set up the serial port on `fd' for 8 bits, no parity,
 * at the requested speed, etc.  If `local' is true, set CLOCAL
 * regardless of whether the modem option was specified.
 */
void
set_up_tty(fd, local)
    int fd, local;
{
    int speed, x, modembits;
    struct termios tios;

    if (tcgetattr(fd, &tios) < 0) {
	syslog(LOG_ERR, "tcgetattr: %m");
	die(1);
    }

    if (!restore_term)
	inittermios = tios;

    tios.c_cflag &= ~(CSIZE | CSTOPB | PARENB | CLOCAL);

    tios.c_cflag |= CS8 | CREAD | HUPCL;
    if (local || !modem)
	tios.c_cflag |= CLOCAL;

    tios.c_iflag = IGNBRK | IGNPAR;
    tios.c_oflag = 0;
    tios.c_lflag = 0;
    tios.c_cc[VMIN] = 1;
    tios.c_cc[VTIME] = 0;

    if (crtscts == -2) {
	tios.c_iflag |= IXON | IXOFF;
	tios.c_cc[VSTOP] = 0x13;	/* DC3 = XOFF = ^S */
	tios.c_cc[VSTART] = 0x11;	/* DC1 = XON  = ^Q */
    }

    speed = translate_speed(inspeed);
    if (speed) {
	cfsetospeed(&tios, speed);
	cfsetispeed(&tios, speed);
    } else {
	speed = cfgetospeed(&tios);
	/*
	 * We can't proceed if the serial port speed is B0,
	 * since that implies that the serial port is disabled.
	 */
	if (speed == B0) {
	    syslog(LOG_ERR, "Baud rate for %s is 0; need explicit baud rate",
		   devnam);
	    die(1);
	}
    }

    if (modem) {
      modembits = TIOCM_RTS | TIOCM_CTS;
      if (ioctl(fd, (crtscts ? TIOCMBIS : TIOCMBIC), &modembits) < 0)
	syslog(LOG_ERR, "ioctl: TIOCMBIS/BIC: %m");
    }

    if (tcsetattr(fd, TCSAFLUSH, &tios) < 0) {
	syslog(LOG_ERR, "tcsetattr: %m");
	die(1);
    }

   baud_rate = inspeed = baud_rate_of(speed);
   restore_term = 1;
}

/*
 * restore_tty - restore the terminal to the saved settings.
 */
void
restore_tty(fd)
    int fd;
{
    if (restore_term) {
	if (tcsetattr(fd, TCSAFLUSH, &inittermios) < 0)
	    if (errno != ENXIO)
		syslog(LOG_WARNING, "tcsetattr: %m");
	restore_term = 0;
    }
}

/*
 * setdtr - control the DTR line on the serial port.
 * This is called from die(), so it shouldn't call die().
 *
 * The write hack is to get NXFax to recognize that there is
 * activity on the port.  Not using the write nukes
 * NXFax's capability to determine port usage.
 *
 */
void
setdtr(fd, on)
int fd, on;
{
    int modembits = TIOCM_DTR;

    if (!on)
      {
	write(fd, " ", 1);
	sleep(1);
      }

/*    ioctl(fd, (on? TIOCMBIS: TIOCMBIC), &modembits); */
    ioctl(fd, (on? TIOCSDTR: TIOCCDTR), 0);
}

/*
 * sifnpmode - Set the mode for handling packets for a given NP.
 */
int
sifnpmode(u, proto, mode)
    int u;
    int proto;
    enum NPmode mode;
{
    struct npioctl npi;

    npi.protocol = proto;
    npi.mode = mode;
    if (ioctl(ppp_fd, PPPIOCSNPMODE, &npi) < 0) {
	syslog(LOG_ERR, "ioctl(set NP %d mode to %d): %m", proto, mode);
	return 0;
    }
    return 1;
}

/*
 * open_ppp_loopback - open the device we use for getting
 * packets in demand mode, and connect it to a ppp interface.
 * Here we use a pty.
 */
void
open_ppp_loopback()
{
    int flags;
    struct termios tios;
    int pppdisc = PPPDISC;

    if (openpty(&loop_master, &loop_slave, loop_name) < 0) {
	syslog(LOG_ERR, "No free pty for loopback");
	die(1);
    }
    SYSDEBUG((LOG_DEBUG, "using %s for loopback", loop_name));

    if (tcgetattr(loop_slave, &tios) == 0) {
	tios.c_cflag &= ~(CSIZE | CSTOPB | PARENB);
	tios.c_cflag |= CS8 | CREAD;
	tios.c_iflag = IGNPAR;
	tios.c_oflag = 0;
	tios.c_lflag = 0;
	if (tcsetattr(loop_slave, TCSAFLUSH, &tios) < 0)
	    syslog(LOG_WARNING, "couldn't set attributes on loopback: %m");
    }

    if ((flags = fcntl(loop_master, F_GETFL)) != -1) 
	if (fcntl(loop_master, F_SETFL, flags | O_NONBLOCK) == -1)
	    syslog(LOG_WARNING, "couldn't set loopback to nonblock: %m");

    ppp_fd = loop_slave;
    if (ioctl(ppp_fd, TIOCSETD, &pppdisc) < 0) {
	syslog(LOG_ERR, "ioctl(TIOCSETD): %m");
	die(1);
    }

    /*
     * Find out which interface we were given.
     */
    if (ioctl(ppp_fd, PPPIOCGUNIT, &ifunit) < 0) {	
	syslog(LOG_ERR, "ioctl(PPPIOCGUNIT): %m");
	die(1);
    }

    /*
     * Enable debug in the driver if requested.
     */
    if (kdebugflag) {
	if (ioctl(ppp_fd, PPPIOCGFLAGS, (caddr_t) &flags) < 0) {
	    syslog(LOG_WARNING, "ioctl (PPPIOCGFLAGS): %m");
	} else {
	    flags |= (kdebugflag & 0xFF) * SC_DEBUG;
	    if (ioctl(ppp_fd, PPPIOCSFLAGS, (caddr_t) &flags) < 0)
		syslog(LOG_WARNING, "ioctl(PPPIOCSFLAGS): %m");
	}
    }
}

/*
 * restore_loop - reattach the ppp unit to the loopback.
 */
void
restore_loop()
{
    int x;

    /*
     * Transfer the ppp interface back to the loopback.
     */
    if (ioctl(ppp_fd, PPPIOCXFERUNIT, 0) < 0) {
	syslog(LOG_ERR, "ioctl(transfer ppp unit in restore_loop()): %m");
//	die(1);
    }
    x = PPPDISC;
    if (ioctl(loop_slave, TIOCSETD, &x) < 0) {
	syslog(LOG_ERR, "ioctl(TIOCSETD): %m");
	die(1);
    }

    /*
     * Check that we got the same unit again.
     */
    if (ioctl(loop_slave, PPPIOCGUNIT, &x) < 0) {	
	syslog(LOG_ERR, "ioctl(PPPIOCGUNIT): %m");
	die(1);
    }
    if (x != ifunit) {
	syslog(LOG_ERR, "transfer_ppp failed: wanted unit %d, got %d",
	       ifunit, x);
	die(1);
    }
    ppp_fd = loop_slave;
}

/*
 * output - Output PPP packet.
 */
void
output(unit, p, len)
    int unit;
    u_char *p;
    int len;
{
    if (debug)
	log_packet(p, len, "sent ", LOG_DEBUG);

    if (write(ttyfd, p, len) < 0) {
	if (errno == EWOULDBLOCK || errno == ENOBUFS
	    || errno == ENXIO || errno == EIO) {
	    syslog(LOG_WARNING, "write: warning: %m");
	} else {
	    syslog(LOG_ERR, "write: %m");
	    die(1);
	}
    }
}

/*
 * wait_input - wait until there is data available on ttyfd,
 * for the length of time specified by *timo (indefinite
 * if timo is NULL).
 */
void
wait_input(timo)
    struct timeval *timo;
{
    fd_set ready;
    int n;

    FD_ZERO(&ready);
    FD_SET(ttyfd, &ready);
    n = select(ttyfd+1, &ready, NULL, &ready, timo);
    if (n < 0 && errno != EINTR) {
	syslog(LOG_ERR, "select: %m");
	die(1);
    }
}

/*
 * wait_loop_output - wait until there is data available on the
 * loopback, for the length of time specified by *timo (indefinite
 * if timo is NULL).
 */
void
wait_loop_output(timo)
    struct timeval *timo;
{
    fd_set ready;
    int n;

    FD_ZERO(&ready);
    FD_SET(loop_master, &ready);
    n = select(loop_master + 1, &ready, NULL, &ready, timo);
    if (n < 0 && errno != EINTR) {
	syslog(LOG_ERR, "select: %m");
	die(1);
    }
}

/*
 * wait_time - wait for a given length of time or until a
 * signal is received.
 */
void
wait_time(timo)
    struct timeval *timo;
{
    int n;

    n = select(0, NULL, NULL, NULL, timo);
    if (n < 0 && errno != EINTR) {
	syslog(LOG_ERR, "select: %m");
	die(1);
    }
}

/*
 * read_packet - get a PPP packet from the serial device.
 */
int
read_packet(buf)
    u_char *buf;
{
    int len;

    if ((len = read(ttyfd, buf, PPP_MTU + PPP_HDRLEN)) < 0) {
	if (errno == EWOULDBLOCK || errno == EINTR) {
	    MAINDEBUG((LOG_DEBUG, "read: %m"));
	    return -1;
	}
	syslog(LOG_ERR, "read: %m");
	die(1);
    }
    return len;
}

/*
 * get_loop_output - read characters from the loopback, form them
 * into frames, and detect when we want to bring the real link up.
 * Return value is 1 if we need to bring up the link, 0 otherwise.
 */
int
get_loop_output()
{
    int rv = 0;
    int n;

    while ((n = read(loop_master, inbuf, sizeof(inbuf))) >= 0) {
	if (loop_chars(inbuf, n))
	    rv = 1;
    }

    if (n == 0) {
	syslog(LOG_ERR, "eof on loopback");
	die(1);
    } else if (errno != EWOULDBLOCK){
	syslog(LOG_ERR, "read from loopback: %m");
//	die(1);
    }

//    return 0;			// sys NeXT orginal
    return rv;
}

/*
 * ppp_send_config - configure the transmit characteristics of
 * the ppp interface.
 */
void
ppp_send_config(unit, mtu, asyncmap, pcomp, accomp)
    int unit, mtu;
    u_int32_t asyncmap;
    int pcomp, accomp;
{
    u_int x;
    struct ifreq ifr;

    strncpy(ifr.ifr_name, ifname, sizeof (ifr.ifr_name));
    ifr.ifr_mtu = mtu;
    if (ioctl(sockfd, SIOCSIFMTU, (caddr_t) &ifr) < 0) {
	syslog(LOG_ERR, "ioctl(SIOCSIFMTU): %m");
	quit();
    }

    if (ioctl(ppp_fd, PPPIOCSASYNCMAP, (caddr_t) &asyncmap) < 0) {
	syslog(LOG_ERR, "ioctl(PPPIOCSASYNCMAP): %m");
	quit();
    }

    if (ioctl(ppp_fd, PPPIOCGFLAGS, (caddr_t) &x) < 0) {
	syslog(LOG_ERR, "ioctl(PPPIOCGFLAGS): %m");
	quit();
    }
    x = pcomp? x | SC_COMP_PROT: x &~ SC_COMP_PROT;
    x = accomp? x | SC_COMP_AC: x &~ SC_COMP_AC;
    if (ioctl(ppp_fd, PPPIOCSFLAGS, (caddr_t) &x) < 0) {
	syslog(LOG_ERR, "ioctl(PPPIOCSFLAGS): %m");
	quit();
    }
}


/*
 * ppp_set_xaccm - set the extended transmit ACCM for the interface.
 */
void
ppp_set_xaccm(unit, accm)
    int unit;
    ext_accm accm;
{
    if (ioctl(ppp_fd, PPPIOCSXASYNCMAP, accm) < 0 && errno != ENOTTY)
	syslog(LOG_WARNING, "ioctl(PPPIOCSXASYNCMAP): %m");
}


/*
 * ppp_recv_config - configure the receive-side characteristics of
 * the ppp interface.
 */
void
ppp_recv_config(unit, mru, asyncmap, pcomp, accomp)
    int unit, mru;
    u_int32_t asyncmap;
    int pcomp, accomp;
{
    int x;

    if (ioctl(ppp_fd, PPPIOCSMRU, (caddr_t) &mru) < 0) {
	syslog(LOG_ERR, "ioctl(PPPIOCSMRU): %m");
	quit();
    }
    if (ioctl(ppp_fd, PPPIOCSRASYNCMAP, (caddr_t) &asyncmap) < 0) {
	syslog(LOG_ERR, "ioctl(PPPIOCSRASYNCMAP): %m");
	quit();
    }
    if (ioctl(ppp_fd, PPPIOCGFLAGS, (caddr_t) &x) < 0) {
	syslog(LOG_ERR, "ioctl(PPPIOCGFLAGS): %m");
	quit();
    }
    x = !accomp? x | SC_REJ_COMP_AC: x &~ SC_REJ_COMP_AC;
    if (ioctl(ppp_fd, PPPIOCSFLAGS, (caddr_t) &x) < 0) {
	syslog(LOG_ERR, "ioctl(PPPIOCSFLAGS): %m");
	quit();
    }
}

/*
 * ccp_test - ask kernel whether a given compression method
 * is acceptable for use.
 */
int
ccp_test(unit, opt_ptr, opt_len, for_transmit)
    int unit, opt_len, for_transmit;
    u_char *opt_ptr;
{
    struct ppp_option_data data;

    data.ptr = opt_ptr;
    data.length = opt_len;
    data.transmit = for_transmit;
    if (ioctl(ttyfd, PPPIOCSCOMPRESS, (caddr_t) &data) >= 0)
	return 1;
    return (errno == ENOBUFS)? 0: -1;
}

/*
 * ccp_flags_set - inform kernel about the current state of CCP.
 */
void
ccp_flags_set(unit, isopen, isup)
    int unit, isopen, isup;
{
    int x;

    if (ioctl(ppp_fd, PPPIOCGFLAGS, (caddr_t) &x) < 0) {
	syslog(LOG_ERR, "ioctl(PPPIOCGFLAGS): %m");
	return;
    }
    x = isopen? x | SC_CCP_OPEN: x &~ SC_CCP_OPEN;
    x = isup? x | SC_CCP_UP: x &~ SC_CCP_UP;
    if (ioctl(ppp_fd, PPPIOCSFLAGS, (caddr_t) &x) < 0)
	syslog(LOG_ERR, "ioctl(PPPIOCSFLAGS): %m");
}

/*
 * ccp_fatal_error - returns 1 if decompression was disabled as a
 * result of an error detected after decompression of a packet,
 * 0 otherwise.  This is necessary because of patent nonsense.
 */
int
ccp_fatal_error(unit)
    int unit;
{
    int x;

    if (ioctl(ppp_fd, PPPIOCGFLAGS, (caddr_t) &x) < 0) {
	syslog(LOG_ERR, "ioctl(PPPIOCGFLAGS): %m");
	return 0;
    }
    return x & SC_DC_FERROR;
}

/*
 * get_idle_time - return how long the link has been idle.
 */
int
get_idle_time(u, ip)
    int u;
    struct ppp_idle *ip;
{
  return (ioctl(ppp_fd, PPPIOCGIDLE, ip) >= 0); 
}

#ifdef PPP_FILTER
/*
 * set_filters - transfer the pass and active filters to the kernel.
 */
int
set_filters(pass, active)
    struct bpf_program *pass, *active;
{
    int ret = 1;

    if (pass->bf_len > 0) {
	if (ioctl(ppp_fd, PPPIOCSPASS, pass) < 0) {
	    syslog(LOG_ERR, "Couldn't set pass-filter in kernel: %m");
	    ret = 0;
	}
    }
    if (active->bf_len > 0) {
	if (ioctl(ppp_fd, PPPIOCSACTIVE, active) < 0) {
	    syslog(LOG_ERR, "Couldn't set active-filter in kernel: %m");
	    ret = 0;
	}
    }
    return ret;
}
#endif

/*
 * sifvjcomp - config tcp header compression
 */
int
sifvjcomp(u, vjcomp, cidcomp, maxcid)
    int u, vjcomp, cidcomp, maxcid;
{
    u_int x;

    if (ioctl(ppp_fd, PPPIOCGFLAGS, (caddr_t) &x) < 0) {
	syslog(LOG_ERR, "ioctl(PPIOCGFLAGS): %m");
	return 0;
    }
    x = vjcomp ? x | SC_COMP_TCP: x &~ SC_COMP_TCP;
    x = cidcomp? x & ~SC_NO_TCP_CCID: x | SC_NO_TCP_CCID;
    if (ioctl(ppp_fd, PPPIOCSFLAGS, (caddr_t) &x) < 0) {
	syslog(LOG_ERR, "ioctl(PPPIOCSFLAGS): %m");
	return 0;
    }
    if (ioctl(ppp_fd, PPPIOCSMAXCID, (caddr_t) &maxcid) < 0) {
	syslog(LOG_ERR, "ioctl(PPPIOCSFLAGS): %m");
	return 0;
    }
    return 1;
}

/*
 * sifup - Config the interface up and enable IP packets to pass.
 */
#ifndef SC_ENABLE_IP
#define SC_ENABLE_IP	0x100	/* compat for old versions of kernel code */
#endif

int
sifup(u)
    int u;
{
    struct ifreq ifr;
    u_int x;
    struct npioctl npi;

    strncpy(ifr.ifr_name, ifname, sizeof (ifr.ifr_name));
    if (ioctl(sockfd, SIOCGIFFLAGS, (caddr_t) &ifr) < 0) {
	syslog(LOG_ERR, "ioctl (SIOCGIFFLAGS): %m");
	return 0;
    }
    ifr.ifr_flags |= IFF_UP;
    if (ioctl(sockfd, SIOCSIFFLAGS, (caddr_t) &ifr) < 0) {
	syslog(LOG_ERR, "ioctl(SIOCSIFFLAGS): %m");
	return 0;
    }
    if_is_up = 1;
    npi.protocol = PPP_IP;
    npi.mode = NPMODE_PASS;
    if (ioctl(ppp_fd, PPPIOCSNPMODE, &npi) < 0) {
	if (errno != ENOTTY) {
	    syslog(LOG_ERR, "ioctl(PPPIOCSNPMODE): %m");
	    return 0;
	}
	/* for backwards compatibility */
	if (ioctl(ppp_fd, PPPIOCGFLAGS, (caddr_t) &x) < 0) {
	    syslog(LOG_ERR, "ioctl (PPPIOCGFLAGS): %m");
	    return 0;
	}
	x |= SC_ENABLE_IP;
	if (ioctl(ppp_fd, PPPIOCSFLAGS, (caddr_t) &x) < 0) {
	    syslog(LOG_ERR, "ioctl(PPPIOCSFLAGS): %m");
	    return 0;
	}
    }
    return 1;
}

/*
 * sifdown - Config the interface down and disable IP.
 */
int
sifdown(u)
    int u;
{
    struct ifreq ifr;
    u_int x;
    int rv;
    struct npioctl npi;

    rv = 1;
    npi.protocol = PPP_IP;
    npi.mode = NPMODE_ERROR;
    ioctl(ppp_fd, PPPIOCSNPMODE, (caddr_t) &npi);
    /* ignore errors, because ppp_fd might have been closed by now. */

    strncpy(ifr.ifr_name, ifname, sizeof (ifr.ifr_name));
    if (ioctl(sockfd, SIOCGIFFLAGS, (caddr_t) &ifr) < 0) {
	syslog(LOG_ERR, "ioctl (SIOCGIFFLAGS): %m");
	rv = 0;
    } else {
	ifr.ifr_flags &= ~IFF_UP;
	if (ioctl(sockfd, SIOCSIFFLAGS, (caddr_t) &ifr) < 0) {
	    syslog(LOG_ERR, "ioctl(SIOCSIFFLAGS): %m");
	    rv = 0;
	} else
	    if_is_up = 0;
    }
    return rv;
}

/*
 * SET_SA_FAMILY - set the sa_family field of a struct sockaddr,
 * if it exists.
 */
#define SET_SA_FAMILY(addr, family)		\
    BZERO((char *) &(addr), sizeof(addr));	\
    addr.sa_family = (family); 

/*
 * sifaddr - Config the interface IP addresses and netmask.
 */
int
sifaddr(u, o, h, m)
    int u;
    u_int32_t o, h, m;
{
    int ret = 1;
    struct ifreq ifr;

    strncpy(ifr.ifr_name, ifname, sizeof(ifr.ifr_name));
    SET_SA_FAMILY(ifr.ifr_addr, AF_INET);
    ((struct sockaddr_in *) &ifr.ifr_addr)->sin_addr.s_addr = o;
    if (ioctl(sockfd, SIOCSIFADDR, (caddr_t) &ifr) < 0) {
	syslog(LOG_ERR, "ioctl(SIOCAIFADDR): %m");
	ret = 0;
    }
    ((struct sockaddr_in *) &ifr.ifr_dstaddr)->sin_addr.s_addr = h;
    if (ioctl(sockfd, SIOCSIFDSTADDR, (caddr_t) &ifr) < 0) {
	syslog(LOG_ERR, "ioctl(SIOCSIFDSTADDR): %m");
	ret = 0;
    }
    if (m != 0) {
	((struct sockaddr_in *) &ifr.ifr_addr)->sin_addr.s_addr = m;
	syslog(LOG_INFO, "Setting interface mask to %s\n", ip_ntoa(m));
	if (ioctl(sockfd, SIOCSIFNETMASK, (caddr_t) &ifr) < 0) {
	    syslog(LOG_ERR, "ioctl(SIOCSIFNETMASK): %m");
	    ret = 0;
	}
    }
    ifaddrs[0] = o;
    ifaddrs[1] = h;
    return ret;
}

/*
 * cifaddr - Clear the interface IP addresses, and delete routes
 * through the interface if possible.
 *
 * N.B.: under NextStep, you can't *delete* an address on an interface,
 * so we change it to 0.0.0.0...  A real hack.  But it simplifies
 * reconnection on the server side.
 */
int
cifaddr(u, o, h)
    int u;
    u_int32_t o, h;
{
    struct rtentry rt;

#if 1
    h = o = 0L;
    (void) sifaddr(u, o, h, 0L);
#endif
    SET_SA_FAMILY(rt.rt_dst, AF_INET);
    ((struct sockaddr_in *) &rt.rt_dst)->sin_addr.s_addr = h;
    SET_SA_FAMILY(rt.rt_gateway, AF_INET);
    ((struct sockaddr_in *) &rt.rt_gateway)->sin_addr.s_addr = o;
    rt.rt_flags = RTF_HOST;
    if (ioctl(sockfd, SIOCDELRT, (caddr_t) &rt) < 0) {
	syslog(LOG_ERR, "ioctl(SIOCDELRT): %m");
	return 0;
    }
    return 1;
}

/*
 * sifdefaultroute - assign a default route through the address given.
 */
int
sifdefaultroute(u, l, g)
    int u;
    u_int32_t l, g;
{
    return dodefaultroute(g, 's');
}

/*
 * cifdefaultroute - delete a default route through the address given.
 */
int
cifdefaultroute(u, l, g)
    int u;
    u_int32_t l, g;
{
    return dodefaultroute(g, 'c');
}

/*
 * dodefaultroute - talk to a routing socket to add/delete a default route.
 */
int
dodefaultroute(g, cmd)
    u_int32_t g;
    int cmd;
{
    static struct rtentry rt;

    SET_SA_FAMILY(rt.rt_dst, AF_INET);
    ((struct sockaddr_in *) &rt.rt_dst)->sin_addr.s_addr = 0L;
    SET_SA_FAMILY(rt.rt_gateway, AF_INET);
    ((struct sockaddr_in *) &rt.rt_gateway)->sin_addr.s_addr = g;
    rt.rt_flags = RTF_GATEWAY;
    if (ioctl(sockfd, (cmd == 's') ? SIOCADDRT : SIOCDELRT, &rt) < 0) {
	syslog(LOG_ERR, "%cifdefaultroute: ioctl(%s): %m", cmd,
	       (cmd == 's') ? "SIOCADDRT" : "SIOCDELRT");
	return 0;
    }
    default_route_gateway = (cmd == 's')? g: 0;
    return 1;
}

/*
 * sifproxyarp - Make a proxy ARP entry for the peer.
 */
int
sifproxyarp(unit, hisaddr)
    int unit;
    u_int32_t hisaddr;
{
    struct arpreq arpreq;

    BZERO(&arpreq, sizeof(arpreq));

    /*
     * Get the hardware address of an interface on the same subnet
     * as our local address.
     */
    if (!get_ether_addr(hisaddr, &arpreq.arp_ha)) {
	syslog(LOG_ERR, "Cannot determine ethernet address for proxy ARP");
	return 0;
    }

    SET_SA_FAMILY(arpreq.arp_pa, AF_INET);
    ((struct sockaddr_in *) &arpreq.arp_pa)->sin_addr.s_addr = hisaddr;
    arpreq.arp_flags = ATF_PERM | ATF_PUBL;
    if (ioctl(sockfd, SIOCSARP, (caddr_t)&arpreq) < 0) {
	syslog(LOG_ERR, "ioctl(SIOCSARP): %m");
	return 0;
    }

    proxy_arp_addr = hisaddr;
    return 1;
}

/*
 * cifproxyarp - Delete the proxy ARP entry for the peer.
 */
int
cifproxyarp(unit, hisaddr)
    int unit;
    u_int32_t hisaddr;
{
    struct arpreq arpreq;

    BZERO(&arpreq, sizeof(arpreq));
    SET_SA_FAMILY(arpreq.arp_pa, AF_INET);
    ((struct sockaddr_in *) &arpreq.arp_pa)->sin_addr.s_addr = hisaddr;
    if (ioctl(sockfd, SIOCDARP, (caddr_t)&arpreq) < 0) {
	syslog(LOG_WARNING, "ioctl(SIOCDARP): %m");
	return 0;
    }
    proxy_arp_addr = 0;
    return 1;
}

/*
 * get_ether_addr - get the hardware address of an interface on the
 * the same subnet as ipaddr.
 */
#define MAX_IFS		32

int
get_ether_addr(ipaddr, hwaddr)
    u_int32_t ipaddr;
    struct sockaddr *hwaddr;
{
    struct ifreq *ifr, *ifend, *ifp;
    u_int32_t ina, mask;
    struct ether_addr dla;
    struct ifreq ifreq;
    struct ifconf ifc;
    struct ifreq ifs[MAX_IFS];
    struct hostent *hostent;

    ifc.ifc_len = sizeof(ifs);
    ifc.ifc_req = ifs;
    if (ioctl(sockfd, SIOCGIFCONF, &ifc) < 0) {
	syslog(LOG_ERR, "ioctl(SIOCGIFCONF): %m");
	return 0;
    }

    /*
     * Scan through looking for an interface with an Internet
     * address on the same subnet as `ipaddr'.
     */
    ifend = (struct ifreq *) (ifc.ifc_buf + ifc.ifc_len);
    for (ifr = ifc.ifc_req; ifr < ifend; ifr = (struct ifreq *)
		((char *)&ifr->ifr_addr + sizeof(struct sockaddr))) {
	if (ifr->ifr_addr.sa_family == AF_INET) {
	    ina = ((struct sockaddr_in *) &ifr->ifr_addr)->sin_addr.s_addr;
	    strncpy(ifreq.ifr_name, ifr->ifr_name, sizeof(ifreq.ifr_name));
	    /*
	     * Check that the interface is up, and not point-to-point
	     * or loopback.
	     */
	    if (ioctl(sockfd, SIOCGIFFLAGS, &ifreq) < 0)
		continue;
	    if ((ifreq.ifr_flags &
		 (IFF_UP|IFF_BROADCAST|IFF_POINTOPOINT|IFF_LOOPBACK|IFF_NOARP))
		 != (IFF_UP|IFF_BROADCAST))
		continue;
	    /*
	     * Get its netmask and check that it's on the right subnet.
	     */
	    if (ioctl(sockfd, SIOCGIFNETMASK, &ifreq) < 0)
		continue;
	    mask = ((struct sockaddr_in *) &ifreq.ifr_addr)->sin_addr.s_addr;
	    if ((ipaddr & mask) != (ina & mask))
		continue;

	    break;
	}
    }

    if (ifr >= ifend)
	return 0;
    syslog(LOG_INFO, "found interface %s for proxy arp", ifr->ifr_name);

    /*
     * Get the hostname and look for an entry using the ethers database.
     * Under NeXTStep this is the best we can do for now.
     */
    if ((hostent = gethostbyaddr((char*)&ina, sizeof(ina), AF_INET)) == NULL)
	return 0;

    if (ether_by_host(hostent->h_name, &dla)) {
	syslog(LOG_INFO, "Add entry for %s in /etc/ethers", hostent->h_name);
	return 0;	/* it's not there */
    }
    hwaddr->sa_family = AF_UNSPEC;
    BCOPY(&dla, hwaddr->sa_data, sizeof(dla));
    return 1;
}

static int
ether_by_host(hostname, etherptr)
    char *hostname;
    struct ether_addr *etherptr;
{
    struct ether_addr *thisptr;
    void *conn;
    ni_id root;
    ni_namelist val;
    char path[256];

    if (!ether_hostton(hostname, etherptr))
	return 0;
    /*
     * We shall now try and
     * find the address in the
     * top domain of netinfo.
     */
    strcat(strcpy(path, "/machines/"), hostname);

    if (ni_open((void *)0, "/", &conn)
     || ni_root(conn, &root)
     || ni_pathsearch(conn, &root, path)
     || ni_lookupprop(conn, &root, "en_address", &val))
	return 1;

    /*
     * Now we can convert the returned string into an ethernet address.
     */
    strcpy(path, val.ni_namelist_val[0]);
    ni_free(conn);
    if ((thisptr = (struct ether_addr*)ether_aton(path)) == NULL)
	return 1;
    BCOPY(thisptr, etherptr, sizeof(struct ether_addr));
    return 0;
}

/*
 * Return user specified netmask, modified by any mask we might determine
 * for address `addr' (in network byte order).
 * Here we scan through the system's list of interfaces, looking for
 * any non-point-to-point interfaces which might appear to be on the same
 * network as `addr'.  If we find any, we OR in their netmask to the
 * user-specified netmask.
 */
u_int32_t
GetMask(addr)
    u_int32_t addr;
{
    u_int32_t mask, nmask, ina;
    struct ifreq *ifr, *ifend, ifreq;
    struct ifconf ifc;
    struct ifreq ifs[MAX_IFS];

    addr = ntohl(addr);
    if (IN_CLASSA(addr))	/* determine network mask for address class */
	nmask = IN_CLASSA_NET;
    else if (IN_CLASSB(addr))
	nmask = IN_CLASSB_NET;
    else
	nmask = IN_CLASSC_NET;
    /* class D nets are disallowed by bad_ip_adrs */
    mask = netmask | htonl(nmask);

    /*
     * Scan through the system's network interfaces.
     */
    ifc.ifc_len = sizeof(ifs);
    ifc.ifc_req = ifs;
    if (ioctl(sockfd, SIOCGIFCONF, &ifc) < 0) {
	syslog(LOG_WARNING, "ioctl(SIOCGIFCONF): %m");
	return mask;
    }
    ifend = (struct ifreq *) (ifc.ifc_buf + ifc.ifc_len);
    for (ifr = ifc.ifc_req; ifr < ifend; ifr = (struct ifreq *)
	 	((char *)&ifr->ifr_addr + sizeof(struct sockaddr))) {
	/*
	 * Check the interface's internet address.
	 */
	if (ifr->ifr_addr.sa_family != AF_INET)
	    continue;
	ina = ((struct sockaddr_in *) &ifr->ifr_addr)->sin_addr.s_addr;
	if ((ntohl(ina) & nmask) != (addr & nmask))
	    continue;
	/*
	 * Check that the interface is up, and not point-to-point or loopback.
	 */
	strncpy(ifreq.ifr_name, ifr->ifr_name, sizeof(ifreq.ifr_name));
	if (ioctl(sockfd, SIOCGIFFLAGS, &ifreq) < 0)
	    continue;
	if ((ifreq.ifr_flags & (IFF_UP|IFF_POINTOPOINT|IFF_LOOPBACK))
	    != IFF_UP)
	    continue;
	/*
	 * Get its netmask and OR it into our mask.
	 */
	if (ioctl(sockfd, SIOCGIFNETMASK, &ifreq) < 0)
	    continue;
	mask |= ((struct sockaddr_in *)&ifreq.ifr_addr)->sin_addr.s_addr;
    }

    return mask;
}

/*
 * lock - create a lock file for the named device
 */
#define LOCK_PREFIX	"/usr/spool/uucp/LCK/LCK.."

int
lock(dev)
    char *dev;
{
    int fd, pid, n;
    char *p;

    if ((p = strrchr(dev, '/')) != NULL)
	dev = p + 1;
    lock_file = malloc(strlen(LOCK_PREFIX) + strlen(dev) + 1);
    if (lock_file == NULL)
	novm("lock file name");
    strcat(strcpy(lock_file, LOCK_PREFIX), dev);

    while ((fd = open(lock_file, O_EXCL | O_CREAT | O_RDWR, 0644)) < 0) {
	if (errno == EEXIST
	    && (fd = open(lock_file, O_RDONLY, 0)) >= 0) {
	    /* Read the lock file to find out who has the device locked */
	    n = read(fd, &pid, sizeof(pid));
	    if (n <= 0) {
		syslog(LOG_ERR, "Can't read pid from lock file %s", lock_file);
		close(fd);
	    } else {
		if (kill(pid, 0) == -1 && errno == ESRCH) {
		    /* pid no longer exists - remove the lock file */
		    if (unlink(lock_file) == 0) {
			close(fd);
			syslog(LOG_NOTICE, "Removed stale lock on %s (pid %d)",
			       dev, pid);
			continue;
		    } else
			syslog(LOG_WARNING, "Couldn't remove stale lock on %s",
			       dev);
		} else
		    syslog(LOG_NOTICE, "Device %s is locked by pid %d",
			   dev, pid);
	    }
	    close(fd);
	} else
	    syslog(LOG_ERR, "Can't create lock file %s: %m", lock_file);
	free(lock_file);
	lock_file = NULL;
	return -1;
    }

    pid = getpid();
    write(fd, &pid, sizeof pid);

    close(fd);
    return 0;
}

/*
 * unlock - remove our lockfile
 */
void
unlock()
{
    if (lock_file) {
	unlink(lock_file);
	free(lock_file);
	lock_file = NULL;
    }
}

/*
 * daemon - Detach us from the terminal session.
 */
int
daemon(nochdir, noclose)
    int nochdir, noclose;
{
    int pid;

    if ((pid = fork()) < 0)
	return -1;
    if (pid != 0)
	exit(0);		/* parent dies */
    (void)setsid();
    if (!nochdir)
	chdir("/");
    if (!noclose) {
	fclose(stdin);		/* don't need stdin, stdout, stderr */
	fclose(stdout);
	fclose(stderr);
    }
    return 0;
}

/*
 * This logwtmp() implementation is subject to the following copyright:
 *
 * Copyright (c) 1988 The Regents of the University of California.
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms are permitted
 * provided that the above copyright notice and this paragraph are
 * duplicated in all such forms and that any documentation,
 * advertising materials, and other materials related to such
 * distribution and use acknowledge that the software was developed
 * by the University of California, Berkeley.  The name of the
 * University may not be used to endorse or promote products derived
 * from this software without specific prior written permission.
 * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR
 * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED
 * WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE.
 */

#define WTMPFILE        "/usr/adm/wtmp"

void
logwtmp(line, name, host)
    const char *line, *name, *host;
{
    int fd;
    struct stat buf;
    struct utmp ut;

    if ((fd = open(WTMPFILE, O_WRONLY|O_APPEND, 0)) < 0)
	return;
    if (!fstat(fd, &buf)) {
	(void)strncpy(ut.ut_line, line, sizeof(ut.ut_line));
	(void)strncpy(ut.ut_name, name, sizeof(ut.ut_name));
	(void)strncpy(ut.ut_host, host, sizeof(ut.ut_host));
	(void)time(&ut.ut_time);
	if (write(fd, (char *)&ut, sizeof(struct utmp)) != sizeof(struct utmp))
	    (void)ftruncate(fd, buf.st_size);
    }
    close(fd);
}


/*
 * note_debug_level - note a change in the debug level.
 */
void
note_debug_level()
{
    if (debug) {
	syslog(LOG_INFO, "Debug turned ON, Level %d", debug);
	setlogmask(LOG_UPTO(LOG_DEBUG));
    } else {
	setlogmask(LOG_UPTO(LOG_WARNING));
    }
}

#if defined(i386) && defined(HAS_BROKEN_IOCTL)
int
ioctl(fd, cmd, c)
    int fd, cmd;
    caddr_t c;
{
#undef	ioctl
    int ret;

#ifdef DEBUGIOCTL
    int serrno;
    u_char let, code, size;

    size = (cmd >> 16) & IOCPARM_MASK;
    let = (cmd >> 8);
    code = cmd;

    if (let == 't' && (75 <= code && code <= 90))
    syslog(LOG_INFO, "ioctl(%d, 0x%x ('%c', %d, %d), 0x%x)\n", fd, cmd,
	   let, code, size, c);
#endif

    ret = ioctl(fd, cmd, c);

#ifdef DEBUGIOCTL
    serrno = errno;
    if (ret == -1)
	syslog(LOG_INFO, "ioctl('%c', %d, %d) errno = %d (%m)\n",
		let, code, size, errno);
    if (let == 't' && (75 <= code && code <= 90) && (cmd & IOC_OUT)) {
	int i, len = ((cmd >> 16) & IOCPARM_MASK);
	for (i = 0; i < len / 4; ++i)
		syslog(LOG_INFO, "word[%d] @ 0x%06x = 0x%x\n",
		       i, &((int *) c)[i],((int *)c)[i]);
    }
    errno = serrno;
#endif

    if (ret == -1 && errno == EPERM)
	errno = ret = 0;
    return ret;
}
#endif	/* HAS_BROKEN_IOCTL */


#if defined(FIXSIGS) && (defined (hppa) || defined(sparc))

/*
 * These redefinitions of Posix functions are necessary
 * because HPPA systems have an OS bug that causes 
 * sigaction to core dump:
 *
 * AlainF 9-Nov-1994	HACK FOR HP-PA/NEXTSTEP
 *			sigaction(3) seems broken in the HP-PA NeXTSTEP 3.2
 *			Posix lib. This causes pppd to SIGBUS at the expiration
 *			of the first timeout (_sigtramp seems to invoke
 *			the SIGALRM handler at an unreasonably low address).
 *			All calls so sigaction(3) have been changed to calls
 *			to sigvec(2) and sigprocmask(SIG_BLOCK,...) to
 *			sigblock(2).
 *			This is kind of a hack, especially since there are
 *			other routines of the Posix lib still used, but
 *			it worked for me.
 *
 * Dave Hess <David-Hess@net.tamu.edu> noted that 3.3 Sparc seems to
 * have the same bug.  Thus this fix has been enabled for SPARC also.
 *
 *
 */

int sigemptyset(sigset_t *mask)
{
  *mask = 0;
}

sigaddset(sigset_t *mask, int which_sig)
{
  *mask |= sigmask(which_sig);
}


int sigaction(int sig, const struct sigaction *act, struct sigaction *oact)
{
   struct sigvec sv;
   static int in = 0;

   sv.sv_handler = act->sa_handler;
   sv.sv_mask = act->sa_mask;
   sv.sv_flags = 0;

   if (!in)
     {
       in = 1;
       syslog(LOG_WARNING, "PPPD: Inside modified HP and SPARC sigaction\n");
     }

   return sigvec(sig, &sv, NULL);
}

#endif


/*
 * Code following is added for 2.3 compatibility
 */


/*
 * List of valid speeds.
 */
struct speed {
    int speed_int, speed_val;
} speeds[] = {
#ifdef B50
    { 50, B50 },
#endif
#ifdef B75
    { 75, B75 },
#endif
#ifdef B110
    { 110, B110 },
#endif
#ifdef B134
    { 134, B134 },
#endif
#ifdef B150
    { 150, B150 },
#endif
#ifdef B200
    { 200, B200 },
#endif
#ifdef B300
    { 300, B300 },
#endif
#ifdef B600
    { 600, B600 },
#endif
#ifdef B1200
    { 1200, B1200 },
#endif
#ifdef B1800
    { 1800, B1800 },
#endif
#ifdef B2000
    { 2000, B2000 },
#endif
#ifdef B2400
    { 2400, B2400 },
#endif
#ifdef B3600
    { 3600, B3600 },
#endif
#ifdef B4800
    { 4800, B4800 },
#endif
#ifdef B7200
    { 7200, B7200 },
#endif
#ifdef B9600
    { 9600, B9600 },
#endif
#ifdef B19200
    { 19200, B19200 },
#endif
#ifdef B38400
    { 38400, B38400 },
#endif
#ifdef EXTA
    { 19200, EXTA },
#endif
#ifdef EXTB
    { 38400, EXTB },
#endif
#ifdef B14400
    { 14400, B14400 },
#endif
#ifdef B28800
    { 28800, B28800 },
#endif
#ifdef B43200
    { 43200, B43200 },
#endif
#ifdef B57600
    { 57600, B57600 },
#endif
/*
#ifndef B115200
#warning Defining B115200
#define B115200 20
#endif
*/
#ifdef B115200
    { 115200, B115200 },
#endif
    { 0, 0 }
};

/*
 * Translate from bits/second to a speed_t.
 */
int
translate_speed(bps)
    int bps;
{
    struct speed *speedp;

    if (bps == 0)
	return 0;
    for (speedp = speeds; speedp->speed_int; speedp++)
	if (bps == speedp->speed_int)
	    return speedp->speed_val;
    syslog(LOG_WARNING, "speed %d not supported", bps);
    return 0;
}

/*
 * Translate from a speed_t to bits/second.
 */
static int
baud_rate_of(speed)
    int speed;
{
    struct speed *speedp;

    if (speed == 0)
	return 0;
    for (speedp = speeds; speedp->speed_int; speedp++)
	if (speed == speedp->speed_val)
	    return speedp->speed_int;
    return 0;
}

char *
strdup(s)
    const char *s;
{
    char *d = malloc(strlen(s) + 1);

    if (d) strcpy(d, s);
    return d;
}
//*****************************************************************************
//
//		attempt to open a pty pair
//
//*****************************************************************************
#define	PTY_TEMPLATE "/dev/pty??"
#define	PTY_LENGTH 11

static int openpty(int *master, int *slave, char *loop_name)
{
char device[PTY_LENGTH];
char *block, *num;
char *blockLoc; // specifies the location of block for the device string
char *numLoc; 	// specifies the pty name with the digit ptyxD
char *msLoc;	// specifies the master (ptyxx) or slave (ttyxx)
struct sgttyb setp = {B9600, B9600, (char)0x7f, (char)0x15, (CRMOD|ANYP)};
struct tchars setc = {CINTR, CQUIT, CSTART, CSTOP, CEOF, CBRK};
struct ltchars sltc = {CSUSP, CDSUSP, CRPRNT, CFLUSH, CWERASE, CLNEXT};
int	lset = (LCRTBS|LCRTERA|LCRTKIL|LCTLECH|LPENDIN|LDECCTQ);
int	setd = NTTYDISC;		/* new tty discipline */
    
    strcpy(device, PTY_TEMPLATE); 	// string constants are not writable
    blockLoc = &device[ strlen("/dev/pty") ];
    numLoc = &device[ strlen("/dev/pty?") ];
    msLoc = &device[ strlen("/dev/") ];
    for (block = "pqrs"; *block; block++)	
		{									
		*blockLoc = *block;
		for (num = "0123456789abcdef"; *num; num++)
			{
			*numLoc = *num;
			*master = open(device, O_RDWR);	// open device, assign fd to master
			if (*master >= 0)				// if master 0 or >, open succesful
				{
				*msLoc = 't';
				*slave = open(device, O_RDWR);
				if (*slave >= 0)
					{
						/* set parameters -- stty */
					(void) ioctl(*slave, TIOCSETP, (char *)&setp);
						/* set special characters */
					(void) ioctl(*slave, TIOCSETC, (char *)&setc);
						/* set line discipline */
					(void) ioctl(*slave, TIOCSETD, (char *)&setd);
						/* set local special chars */
					(void) ioctl(*slave, TIOCSLTC, (char *)&sltc);
						/* set entire local mode word */
					(void) ioctl(*slave, TIOCLSET, (char *)&lset);
					strncpy(loop_name, device, PTY_LENGTH);		// global var
					
					return 1;
					}
				}
			} /* hunting through a bank of ptys */
		} /* hunting through blocks of ptys in all the right places */
		
    *master = -1;		//If we get here we were unable to open 
    *slave = -1;		//master and/or slave pty's 

	return -1;
}
//*****************************************************************************
//
// 		cleanup_defaultroute 
//
//*****************************************************************************

void
cleanup_defaultroute()
{
static u_int32_t ina, outa;
static struct ifreq ifr;
static int setd = NTTYDISC;		/* new tty discipline */


    if (default_route_gateway)
	cifdefaultroute(0, 0, default_route_gateway);

							// load ifreq struct
    strncpy(ifr.ifr_name, ifname, sizeof(ifr.ifr_name));
    SET_SA_FAMILY(ifr.ifr_addr, AF_INET);

							// delete route
    if (ioctl(sockfd, SIOCSIFADDR, (caddr_t) &ifr) < 0) 
		syslog(LOG_ERR, "ioctl(SIOCAIFADDR): %m");
	else
	    ina = ((struct sockaddr_in *) &ifr.ifr_addr)->sin_addr.s_addr;
    if (ioctl(sockfd, SIOCSIFDSTADDR, (caddr_t) &ifr) < 0) 
		syslog(LOG_ERR, "ioctl(SIOCSIFDSTADDR): %m");
	else
   		outa = ((struct sockaddr_in *) &ifr.ifr_dstaddr)->sin_addr.s_addr;
    cifaddr(0, ina, outa);		 
							// set interface flag down
    if (ioctl(sockfd, SIOCGIFFLAGS, (caddr_t) &ifr) < 0) 	
		syslog(LOG_ERR, "ioctl (SIOCGIFFLAGS): %m");
    else
		{
		ifr.ifr_flags &= ~IFF_UP;
		if (ioctl(sockfd, SIOCSIFFLAGS, (caddr_t) &ifr) < 0) 
	    	syslog(LOG_ERR, "ioctl(SIOCSIFFLAGS): %m");
		}

			// restore pty slave to ttydisc so that we can get ppp0 for serial
    if (ioctl(ppp_fd, TIOCSETD, &setd) < 0) {
	syslog(LOG_ERR, "ioctl(TIOCSETD): %m");
	die(1);
    }

}

void
if_stats()
{
    static struct npioctl npi;
	static int flags;

    npi.protocol = PPP_IP;
    npi.mode = NPMODE_ERROR;
    if (ioctl(ppp_fd, PPPIOCGNPMODE, &npi) < 0) 
		syslog(LOG_ERR, "ioctl sys stats: %m");
	if (ioctl(ppp_fd, PPPIOCGFLAGS, (caddr_t) &flags) < 0)
		syslog(LOG_WARNING, "ioctl(PPPIOCSFLAGS): %m");
}

port_t
gk_report(status)
	int status;
{
port_t ret;

	gk_state = status;
	if(gk_server == PORT_NULL)				// Find the GateKeeper server.
		{
		if((gk_server = gk_rpc_server_look_up()) == PORT_NULL)
			syslog(LOG_ERR, "Unable to find GateKeeper RPC server");
		}
											// inform GateKeeper of our state
	if((ret = set_gk(gk_server, status)) != KERN_SUCCESS) 			 
		syslog(LOG_ERR, "Call to GateKeeper server failed.");

    return ret;
}

void
gk_report_oneway(status)
	int status;
{
	gk_state = status;
	if(gk_server == PORT_NULL)				// Find the GateKeeper server.
		{
		if((gk_server = gk_rpc_server_look_up()) == PORT_NULL)
			syslog(LOG_ERR, "Unable to find GateKeeper RPC server");
		}
											
	set_gk_oneway(gk_server, status); 		// inform GateKeeper of our state
}

These are the contents of the former NiCE NeXT User Group NeXTSTEP/OpenStep software archive, currently hosted by Netfuture.ch.