src/dhcp.c

/* dnsmasq is Copyright (c) 2000-2015 Simon Kelley

   This program is free software; you can redistribute it and/or modify
   it under the terms of the GNU General Public License as published by
   the Free Software Foundation; version 2 dated June, 1991, or
   (at your option) version 3 dated 29 June, 2007.
 
   This program is distributed in the hope that it will be useful,
   but WITHOUT ANY WARRANTY; without even the implied warranty of
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
   GNU General Public License for more details.
     
   You should have received a copy of the GNU General Public License
   along with this program.  If not, see <http://www.gnu.org/licenses/>.
*/

#include "dnsmasq.h"

#ifdef HAVE_DHCP

struct iface_param {
  struct dhcp_context *current;
  struct dhcp_relay *relay;
  struct in_addr relay_local;
  int ind;
};

struct match_param {
  int ind, matched;
  struct in_addr netmask, broadcast, addr;
};

static int complete_context(struct in_addr local, int if_index, char *label,
			    struct in_addr netmask, struct in_addr broadcast, void *vparam);
static int check_listen_addrs(struct in_addr local, int if_index, char *label,
			      struct in_addr netmask, struct in_addr broadcast, void *vparam);
static int relay_upstream4(struct dhcp_relay *relay, struct dhcp_packet *mess, size_t sz, int iface_index);
static struct dhcp_relay *relay_reply4(struct dhcp_packet *mess, char *arrival_interface);

static int make_fd(int port)
{
  int fd = socket(PF_INET, SOCK_DGRAM, IPPROTO_UDP);
  struct sockaddr_in saddr;
  int oneopt = 1;
#if defined(IP_MTU_DISCOVER) && defined(IP_PMTUDISC_DONT)
  int mtu = IP_PMTUDISC_DONT;
#endif
#if defined(IP_TOS) && defined(IPTOS_CLASS_CS6)
  int tos = IPTOS_CLASS_CS6;
#endif

  if (fd == -1)
    die (_("cannot create DHCP socket: %s"), NULL, EC_BADNET);
  
  if (!fix_fd(fd) ||
#if defined(IP_MTU_DISCOVER) && defined(IP_PMTUDISC_DONT)
      setsockopt(fd, IPPROTO_IP, IP_MTU_DISCOVER, &mtu, sizeof(mtu)) == -1 ||
#endif
#if defined(IP_TOS) && defined(IPTOS_CLASS_CS6)
      setsockopt(fd, IPPROTO_IP, IP_TOS, &tos, sizeof(tos)) == -1 ||
#endif
#if defined(HAVE_LINUX_NETWORK)
      setsockopt(fd, IPPROTO_IP, IP_PKTINFO, &oneopt, sizeof(oneopt)) == -1 ||
#else
      setsockopt(fd, IPPROTO_IP, IP_RECVIF, &oneopt, sizeof(oneopt)) == -1 ||
#endif
      setsockopt(fd, SOL_SOCKET, SO_BROADCAST, &oneopt, sizeof(oneopt)) == -1)  
    die(_("failed to set options on DHCP socket: %s"), NULL, EC_BADNET);
  
  /* When bind-interfaces is set, there might be more than one dnmsasq
     instance binding port 67. That's OK if they serve different networks.
     Need to set REUSEADDR|REUSEPORT to make this posible.
     Handle the case that REUSEPORT is defined, but the kernel doesn't 
     support it. This handles the introduction of REUSEPORT on Linux. */
  if (option_bool(OPT_NOWILD) || option_bool(OPT_CLEVERBIND))
    {
      int rc = 0;

#ifdef SO_REUSEPORT
      if ((rc = setsockopt(fd, SOL_SOCKET, SO_REUSEPORT, &oneopt, sizeof(oneopt))) == -1 && 
	  errno == ENOPROTOOPT)
	rc = 0;
#endif
      
      if (rc != -1)
	rc = setsockopt(fd, SOL_SOCKET, SO_REUSEADDR, &oneopt, sizeof(oneopt));
      
      if (rc == -1)
	die(_("failed to set SO_REUSE{ADDR|PORT} on DHCP socket: %s"), NULL, EC_BADNET);
    }
  
  memset(&saddr, 0, sizeof(saddr));
  saddr.sin_family = AF_INET;
  saddr.sin_port = htons(port);
  saddr.sin_addr.s_addr = INADDR_ANY;
#ifdef HAVE_SOCKADDR_SA_LEN
  saddr.sin_len = sizeof(struct sockaddr_in);
#endif

  if (bind(fd, (struct sockaddr *)&saddr, sizeof(struct sockaddr_in)))
    die(_("failed to bind DHCP server socket: %s"), NULL, EC_BADNET);

  return fd;
}

void dhcp_init(void)
{
#if defined(HAVE_BSD_NETWORK)
  int oneopt = 1;
#endif

  daemon->dhcpfd = make_fd(daemon->dhcp_server_port);
  if (daemon->enable_pxe)
    daemon->pxefd = make_fd(PXE_PORT);
  else
    daemon->pxefd = -1;

#if defined(HAVE_BSD_NETWORK)
  /* When we're not using capabilities, we need to do this here before
     we drop root. Also, set buffer size small, to avoid wasting
     kernel buffers */
  
  if (option_bool(OPT_NO_PING))
    daemon->dhcp_icmp_fd = -1;
  else if ((daemon->dhcp_icmp_fd = make_icmp_sock()) == -1 ||
	   setsockopt(daemon->dhcp_icmp_fd, SOL_SOCKET, SO_RCVBUF, &oneopt, sizeof(oneopt)) == -1 )
    die(_("cannot create ICMP raw socket: %s."), NULL, EC_BADNET);
  
  /* Make BPF raw send socket */
  init_bpf();
#endif  
}

void dhcp_packet(time_t now, int pxe_fd)
{
  int fd = pxe_fd ? daemon->pxefd : daemon->dhcpfd;
  struct dhcp_packet *mess;
  struct dhcp_context *context;
  struct dhcp_relay *relay;
  int is_relay_reply = 0;
  struct iname *tmp;
  struct ifreq ifr;
  struct msghdr msg;
  struct sockaddr_in dest;
  struct cmsghdr *cmptr;
  struct iovec iov;
  ssize_t sz; 
  int iface_index = 0, unicast_dest = 0, is_inform = 0;
  struct in_addr iface_addr;
  struct iface_param parm;
#ifdef HAVE_LINUX_NETWORK
  struct arpreq arp_req;
#endif
  
  union {
    struct cmsghdr align; /* this ensures alignment */
#if defined(HAVE_LINUX_NETWORK)
    char control[CMSG_SPACE(sizeof(struct in_pktinfo))];
#elif defined(HAVE_SOLARIS_NETWORK)
    char control[CMSG_SPACE(sizeof(unsigned int))];
#elif defined(HAVE_BSD_NETWORK) 
    char control[CMSG_SPACE(sizeof(struct sockaddr_dl))];
#endif
  } control_u;
  struct dhcp_bridge *bridge, *alias;

  msg.msg_controllen = sizeof(control_u);
  msg.msg_control = control_u.control;
  msg.msg_name = &dest;
  msg.msg_namelen = sizeof(dest);
  msg.msg_iov = &daemon->dhcp_packet;
  msg.msg_iovlen = 1;
  
  if ((sz = recv_dhcp_packet(fd, &msg)) == -1 || 
      (sz < (ssize_t)(sizeof(*mess) - sizeof(mess->options)))) 
    return;
    
  #if defined (HAVE_LINUX_NETWORK)
  if (msg.msg_controllen >= sizeof(struct cmsghdr))
    for (cmptr = CMSG_FIRSTHDR(&msg); cmptr; cmptr = CMSG_NXTHDR(&msg, cmptr))
      if (cmptr->cmsg_level == IPPROTO_IP && cmptr->cmsg_type == IP_PKTINFO)
	{
	  union {
	    unsigned char *c;
	    struct in_pktinfo *p;
	  } p;
	  p.c = CMSG_DATA(cmptr);
	  iface_index = p.p->ipi_ifindex;
	  if (p.p->ipi_addr.s_addr != INADDR_BROADCAST)
	    unicast_dest = 1;
	}

#elif defined(HAVE_BSD_NETWORK) 
  if (msg.msg_controllen >= sizeof(struct cmsghdr))
    for (cmptr = CMSG_FIRSTHDR(&msg); cmptr; cmptr = CMSG_NXTHDR(&msg, cmptr))
      if (cmptr->cmsg_level == IPPROTO_IP && cmptr->cmsg_type == IP_RECVIF)
        {
	  union {
            unsigned char *c;
            struct sockaddr_dl *s;
          } p;
	  p.c = CMSG_DATA(cmptr);
	  iface_index = p.s->sdl_index;
	}
  
#elif defined(HAVE_SOLARIS_NETWORK) 
  if (msg.msg_controllen >= sizeof(struct cmsghdr))
    for (cmptr = CMSG_FIRSTHDR(&msg); cmptr; cmptr = CMSG_NXTHDR(&msg, cmptr))
      if (cmptr->cmsg_level == IPPROTO_IP && cmptr->cmsg_type == IP_RECVIF)
	{
	  union {
	    unsigned char *c;
	    unsigned int *i;
	  } p;
	  p.c = CMSG_DATA(cmptr);
	  iface_index = *(p.i);
	}
#endif
	
  if (!indextoname(daemon->dhcpfd, iface_index, ifr.ifr_name))
    return;

#ifdef HAVE_LINUX_NETWORK
  /* ARP fiddling uses original interface even if we pretend to use a different one. */
  strncpy(arp_req.arp_dev, ifr.ifr_name, 16);
#endif 

  /* If the interface on which the DHCP request was received is an
     alias of some other interface (as specified by the
     --bridge-interface option), change ifr.ifr_name so that we look
     for DHCP contexts associated with the aliased interface instead
     of with the aliasing one. */
  for (bridge = daemon->bridges; bridge; bridge = bridge->next)
    {
      for (alias = bridge->alias; alias; alias = alias->next)
	if (wildcard_matchn(alias->iface, ifr.ifr_name, IF_NAMESIZE))
	  {
	    if (!(iface_index = if_nametoindex(bridge->iface)))
	      {
		my_syslog(MS_DHCP | LOG_WARNING,
			  _("unknown interface %s in bridge-interface"),
			  bridge->iface);
		return;
	      }
	    else 
	      {
		strncpy(ifr.ifr_name,  bridge->iface, IF_NAMESIZE);
		break;
	      }
	  }
      
      if (alias)
	break;
    }

#ifdef MSG_BCAST
  /* OpenBSD tells us when a packet was broadcast */
  if (!(msg.msg_flags & MSG_BCAST))
    unicast_dest = 1;
#endif
  
  if ((relay = relay_reply4((struct dhcp_packet *)daemon->dhcp_packet.iov_base, ifr.ifr_name)))
    {
      /* Reply from server, using us as relay. */
      iface_index = relay->iface_index;
      if (!indextoname(daemon->dhcpfd, iface_index, ifr.ifr_name))
	return;
      is_relay_reply = 1; 
      iov.iov_len = sz;
#ifdef HAVE_LINUX_NETWORK
      strncpy(arp_req.arp_dev, ifr.ifr_name, 16);
#endif 
    }
  else
    {
      ifr.ifr_addr.sa_family = AF_INET;
      if (ioctl(daemon->dhcpfd, SIOCGIFADDR, &ifr) != -1 )
	iface_addr = ((struct sockaddr_in *) &ifr.ifr_addr)->sin_addr;
      else
	{
	  my_syslog(MS_DHCP | LOG_WARNING, _("DHCP packet received on %s which has no address"), ifr.ifr_name);
	  return;
	}
      
      for (tmp = daemon->dhcp_except; tmp; tmp = tmp->next)
	if (tmp->name && wildcard_match(tmp->name, ifr.ifr_name))
	  return;
      
      /* unlinked contexts/relays are marked by context->current == context */
      for (context = daemon->dhcp; context; context = context->next)
	context->current = context;
      
      for (relay = daemon->relay4; relay; relay = relay->next)
	relay->current = relay;
      
      parm.current = NULL;
      parm.relay = NULL;
      parm.relay_local.s_addr = 0;
      parm.ind = iface_index;
      
      if (!iface_check(AF_INET, (struct all_addr *)&iface_addr, ifr.ifr_name, NULL))
	{
	  /* If we failed to match the primary address of the interface, see if we've got a --listen-address
	     for a secondary */
	  struct match_param match;
	  
	  match.matched = 0;
	  match.ind = iface_index;
	  
	  if (!daemon->if_addrs ||
	      !iface_enumerate(AF_INET, &match, check_listen_addrs) ||
	      !match.matched)
	    return;
	  
	  iface_addr = match.addr;
	  /* make sure secondary address gets priority in case
	     there is more than one address on the interface in the same subnet */
	  complete_context(match.addr, iface_index, NULL, match.netmask, match.broadcast, &parm);
	}    
      
      if (!iface_enumerate(AF_INET, &parm, complete_context))
	return;

      /* We're relaying this request */
      if  (parm.relay_local.s_addr != 0 &&
	   relay_upstream4(parm.relay, (struct dhcp_packet *)daemon->dhcp_packet.iov_base, (size_t)sz, iface_index))
	return;

      /* May have configured relay, but not DHCP server */
      if (!daemon->dhcp)
	return;

      lease_prune(NULL, now); /* lose any expired leases */
      iov.iov_len = dhcp_reply(parm.current, ifr.ifr_name, iface_index, (size_t)sz, 
			       now, unicast_dest, &is_inform, pxe_fd, iface_addr);
      lease_update_file(now);
      lease_update_dns(0);
      
      if (iov.iov_len == 0)
	return;
    }
  
  msg.msg_name = &dest;
  msg.msg_namelen = sizeof(dest);
  msg.msg_control = NULL;
  msg.msg_controllen = 0;
  msg.msg_iov = &iov;
  iov.iov_base = daemon->dhcp_packet.iov_base;
  
  /* packet buffer may have moved */
  mess = (struct dhcp_packet *)daemon->dhcp_packet.iov_base;
  
#ifdef HAVE_SOCKADDR_SA_LEN
  dest.sin_len = sizeof(struct sockaddr_in);
#endif
  
  if (pxe_fd)
    { 
      if (mess->ciaddr.s_addr != 0)
	dest.sin_addr = mess->ciaddr;
    }
  else if (mess->giaddr.s_addr && !is_relay_reply)
    {
      /* Send to BOOTP relay  */
      dest.sin_port = htons(daemon->dhcp_server_port);
      dest.sin_addr = mess->giaddr; 
    }
  else if (mess->ciaddr.s_addr)
    {
      /* If the client's idea of its own address tallys with
	 the source address in the request packet, we believe the
	 source port too, and send back to that.  If we're replying 
	 to a DHCPINFORM, trust the source address always. */
      if ((!is_inform && dest.sin_addr.s_addr != mess->ciaddr.s_addr) ||
	  dest.sin_port == 0 || dest.sin_addr.s_addr == 0 || is_relay_reply)
	{
	  dest.sin_port = htons(daemon->dhcp_client_port); 
	  dest.sin_addr = mess->ciaddr;
	}
    } 
#if defined(HAVE_LINUX_NETWORK)
  else
    {
      /* fill cmsg for outbound interface (both broadcast & unicast) */
      struct in_pktinfo *pkt;
      msg.msg_control = control_u.control;
      msg.msg_controllen = sizeof(control_u);
      cmptr = CMSG_FIRSTHDR(&msg);
      pkt = (struct in_pktinfo *)CMSG_DATA(cmptr);
      pkt->ipi_ifindex = iface_index;
      pkt->ipi_spec_dst.s_addr = 0;
      msg.msg_controllen = cmptr->cmsg_len = CMSG_LEN(sizeof(struct in_pktinfo));
      cmptr->cmsg_level = IPPROTO_IP;
      cmptr->cmsg_type = IP_PKTINFO;

      if ((ntohs(mess->flags) & 0x8000) || mess->hlen == 0 ||
         mess->hlen > sizeof(ifr.ifr_addr.sa_data) || mess->htype == 0)
        {
          /* broadcast to 255.255.255.255 (or mac address invalid) */
          dest.sin_addr.s_addr = INADDR_BROADCAST;
          dest.sin_port = htons(daemon->dhcp_client_port);
        }
      else
        {
          /* unicast to unconfigured client. Inject mac address direct into ARP cache.
          struct sockaddr limits size to 14 bytes. */
          dest.sin_addr = mess->yiaddr;
          dest.sin_port = htons(daemon->dhcp_client_port);
          memcpy(&arp_req.arp_pa, &dest, sizeof(struct sockaddr_in));
          arp_req.arp_ha.sa_family = mess->htype;
          memcpy(arp_req.arp_ha.sa_data, mess->chaddr, mess->hlen);
          /* interface name already copied in */
          arp_req.arp_flags = ATF_COM;
          if (ioctl(daemon->dhcpfd, SIOCSARP, &arp_req) == -1)
            my_syslog(MS_DHCP | LOG_ERR, _("ARP-cache injection failed: %s"), strerror(errno));
        }
    }
#elif defined(HAVE_SOLARIS_NETWORK)
  else if ((ntohs(mess->flags) & 0x8000) || mess->hlen != ETHER_ADDR_LEN || mess->htype != ARPHRD_ETHER)
    {
      /* broadcast to 255.255.255.255 (or mac address invalid) */
      dest.sin_addr.s_addr = INADDR_BROADCAST;
      dest.sin_port = htons(daemon->dhcp_client_port);
      /* note that we don't specify the interface here: that's done by the
	 IP_BOUND_IF sockopt lower down. */
    }
  else
    {
      /* unicast to unconfigured client. Inject mac address direct into ARP cache. 
	 Note that this only works for ethernet on solaris, because we use SIOCSARP
	 and not SIOCSXARP, which would be perfect, except that it returns ENXIO 
	 mysteriously. Bah. Fall back to broadcast for other net types. */
      struct arpreq req;
      dest.sin_addr = mess->yiaddr;
      dest.sin_port = htons(daemon->dhcp_client_port);
      *((struct sockaddr_in *)&req.arp_pa) = dest;
      req.arp_ha.sa_family = AF_UNSPEC;
      memcpy(req.arp_ha.sa_data, mess->chaddr, mess->hlen);
      req.arp_flags = ATF_COM;
      ioctl(daemon->dhcpfd, SIOCSARP, &req);
    }
#elif defined(HAVE_BSD_NETWORK)
  else 
    {
      send_via_bpf(mess, iov.iov_len, iface_addr, &ifr);
      return;
    }
#endif
   
#ifdef HAVE_SOLARIS_NETWORK
  setsockopt(fd, IPPROTO_IP, IP_BOUND_IF, &iface_index, sizeof(iface_index));
#endif
  
  while(retry_send(sendmsg(fd, &msg, 0)));

  /* This can fail when, eg, iptables DROPS destination 255.255.255.255 */
  if (errno != 0)
    my_syslog(MS_DHCP | LOG_WARNING, _("Error sending DHCP packet to %s: %s"),
	      inet_ntoa(dest.sin_addr), strerror(errno));
}

/* check against secondary interface addresses */
static int check_listen_addrs(struct in_addr local, int if_index, char *label,
			      struct in_addr netmask, struct in_addr broadcast, void *vparam)
{
  struct match_param *param = vparam;
  struct iname *tmp;

  (void) label;

  if (if_index == param->ind)
    {
      for (tmp = daemon->if_addrs; tmp; tmp = tmp->next)
	if ( tmp->addr.sa.sa_family == AF_INET &&
	     tmp->addr.in.sin_addr.s_addr == local.s_addr)
	  {
	    param->matched = 1;
	    param->addr = local;
	    param->netmask = netmask;
	    param->broadcast = broadcast;
	    break;
	  }
    }
  
  return 1;
}

/* This is a complex routine: it gets called with each (address,netmask,broadcast) triple 
   of each interface (and any relay address) and does the  following things:

   1) Discards stuff for interfaces other than the one on which a DHCP packet just arrived.
   2) Fills in any netmask and broadcast addresses which have not been explicitly configured.
   3) Fills in local (this host) and router (this host or relay) addresses.
   4) Links contexts which are valid for hosts directly connected to the arrival interface on ->current.

   Note that the current chain may be superceded later for configured hosts or those coming via gateways. */

static int complete_context(struct in_addr local, int if_index, char *label,
			    struct in_addr netmask, struct in_addr broadcast, void *vparam)
{
  struct dhcp_context *context;
  struct dhcp_relay *relay;
  struct iface_param *param = vparam;

  (void)label;
  
  for (context = daemon->dhcp; context; context = context->next)
    {
      if (!(context->flags & CONTEXT_NETMASK) &&
	  (is_same_net(local, context->start, netmask) ||
	   is_same_net(local, context->end, netmask)))
      { 
	if (context->netmask.s_addr != netmask.s_addr &&
	    !(is_same_net(local, context->start, netmask) &&
	      is_same_net(local, context->end, netmask)))
	  {
	    strcpy(daemon->dhcp_buff, inet_ntoa(context->start));
	    strcpy(daemon->dhcp_buff2, inet_ntoa(context->end));
	    my_syslog(MS_DHCP | LOG_WARNING, _("DHCP range %s -- %s is not consistent with netmask %s"),
		      daemon->dhcp_buff, daemon->dhcp_buff2, inet_ntoa(netmask));
	  }	
 	context->netmask = netmask;
      }
      
      if (context->netmask.s_addr != 0 &&
	  is_same_net(local, context->start, context->netmask) &&
	  is_same_net(local, context->end, context->netmask))
	{
	  /* link it onto the current chain if we've not seen it before */
	  if (if_index == param->ind && context->current == context)
	    {
	      context->router = local;
	      context->local = local;
	      context->current = param->current;
	      param->current = context;
	    }
	  
	  if (!(context->flags & CONTEXT_BRDCAST))
	    {
	      if (is_same_net(broadcast, context->start, context->netmask))
		context->broadcast = broadcast;
	      else 
		context->broadcast.s_addr  = context->start.s_addr | ~context->netmask.s_addr;
	    }
	}		
    }

  for (relay = daemon->relay4; relay; relay = relay->next)
    if (if_index == param->ind && relay->local.addr.addr4.s_addr == local.s_addr && relay->current == relay &&
	(param->relay_local.s_addr == 0 || param->relay_local.s_addr == local.s_addr))
      {
	relay->current = param->relay;
	param->relay = relay;
	param->relay_local = local;	
      }

  return 1;
}
	  
struct dhcp_context *address_available(struct dhcp_context *context, 
				       struct in_addr taddr,
				       struct dhcp_netid *netids)
{
  /* Check is an address is OK for this network, check all
     possible ranges. Make sure that the address isn't in use
     by the server itself. */
  
  unsigned int start, end, addr = ntohl(taddr.s_addr);
  struct dhcp_context *tmp;

  for (tmp = context; tmp; tmp = tmp->current)
    if (taddr.s_addr == context->router.s_addr)
      return NULL;
  
  for (tmp = context; tmp; tmp = tmp->current)
    {
      start = ntohl(tmp->start.s_addr);
      end = ntohl(tmp->end.s_addr);

      if (!(tmp->flags & (CONTEXT_STATIC | CONTEXT_PROXY)) &&
	  addr >= start &&
	  addr <= end &&
	  match_netid(tmp->filter, netids, 1))
	return tmp;
    }

  return NULL;
}

struct dhcp_context *narrow_context(struct dhcp_context *context, 
				    struct in_addr taddr,
				    struct dhcp_netid *netids)
{
  /* We start of with a set of possible contexts, all on the current physical interface.
     These are chained on ->current.
     Here we have an address, and return the actual context correponding to that
     address. Note that none may fit, if the address came a dhcp-host and is outside
     any dhcp-range. In that case we return a static range if possible, or failing that,
     any context on the correct subnet. (If there's more than one, this is a dodgy 
     configuration: maybe there should be a warning.) */
  
  struct dhcp_context *tmp;

  if (!(tmp = address_available(context, taddr, netids)))
    {
      for (tmp = context; tmp; tmp = tmp->current)
	if (match_netid(tmp->filter, netids, 1) &&
	    is_same_net(taddr, tmp->start, tmp->netmask) && 
	    (tmp->flags & CONTEXT_STATIC))
	  break;
      
      if (!tmp)
	for (tmp = context; tmp; tmp = tmp->current)
	  if (match_netid(tmp->filter, netids, 1) &&
	      is_same_net(taddr, tmp->start, tmp->netmask) &&
	      !(tmp->flags & CONTEXT_PROXY))
	    break;
    }
  
  /* Only one context allowed now */
  if (tmp)
    tmp->current = NULL;
  
  return tmp;
}

struct dhcp_config *config_find_by_address(struct dhcp_config *configs, struct in_addr addr)
{
  struct dhcp_config *config;
  
  for (config = configs; config; config = config->next)
    if ((config->flags & CONFIG_ADDR) && config->addr.s_addr == addr.s_addr)
      return config;

  return NULL;
}

int address_allocate(struct dhcp_context *context,
		     struct in_addr *addrp, unsigned char *hwaddr, int hw_len, 
		     struct dhcp_netid *netids, time_t now)   
{
  /* Find a free address: exclude anything in use and anything allocated to
     a particular hwaddr/clientid/hostname in our configuration.
     Try to return from contexts which match netids first. */

  struct in_addr start, addr;
  struct dhcp_context *c, *d;
  int i, pass;
  unsigned int j; 

  /* hash hwaddr: use the SDBM hashing algorithm.  Seems to give good
     dispersal even with similarly-valued "strings". */ 
  for (j = 0, i = 0; i < hw_len; i++)
    j += hwaddr[i] + (j << 6) + (j << 16) - j;
  
  for (pass = 0; pass <= 1; pass++)
    for (c = context; c; c = c->current)
      if (c->flags & (CONTEXT_STATIC | CONTEXT_PROXY))
	continue;
      else if (!match_netid(c->filter, netids, pass))
	continue;
      else
	{
	  if (option_bool(OPT_CONSEC_ADDR))
	    /* seed is largest extant lease addr in this context */
	    start = lease_find_max_addr(c);
	  else
	    /* pick a seed based on hwaddr */
	    start.s_addr = htonl(ntohl(c->start.s_addr) + 
				 ((j + c->addr_epoch) % (1 + ntohl(c->end.s_addr) - ntohl(c->start.s_addr))));

	  /* iterate until we find a free address. */
	  addr = start;
	  
	  do {
	    /* eliminate addresses in use by the server. */
	    for (d = context; d; d = d->current)
	      if (addr.s_addr == d->router.s_addr)
		break;

	    /* Addresses which end in .255 and .0 are broken in Windows even when using 
	       supernetting. ie dhcp-range=192.168.0.1,192.168.1.254,255,255,254.0
	       then 192.168.0.255 is a valid IP address, but not for Windows as it's
	       in the class C range. See  KB281579. We therefore don't allocate these 
	       addresses to avoid hard-to-diagnose problems. Thanks Bill. */	    
	    if (!d &&
		!lease_find_by_addr(addr) && 
		!config_find_by_address(daemon->dhcp_conf, addr) &&
		(!IN_CLASSC(ntohl(addr.s_addr)) || 
		 ((ntohl(addr.s_addr) & 0xff) != 0xff && ((ntohl(addr.s_addr) & 0xff) != 0x0))))
	      {
		struct ping_result *r, *victim = NULL;
		int count, max = (int)(0.6 * (((float)PING_CACHE_TIME)/
					      ((float)PING_WAIT)));
		
		*addrp = addr;

		/* check if we failed to ping addr sometime in the last
		   PING_CACHE_TIME seconds. If so, assume the same situation still exists.
		   This avoids problems when a stupid client bangs
		   on us repeatedly. As a final check, if we did more
		   than 60% of the possible ping checks in the last 
		   PING_CACHE_TIME, we are in high-load mode, so don't do any more. */
		for (count = 0, r = daemon->ping_results; r; r = r->next)
		  if (difftime(now, r->time) >  (float)PING_CACHE_TIME)
		    victim = r; /* old record */
		  else 
		    {
		      count++;
		      if (r->addr.s_addr == addr.s_addr)
			{
			  /* consec-ip mode: we offered this address for another client
			     (different hash) recently, don't offer it to this one. */
			  if (option_bool(OPT_CONSEC_ADDR) && r->hash != j)
			    break;
			  
			  return 1;
			}
		    }

		if (!r) 
		  {
		    if ((count < max) && !option_bool(OPT_NO_PING) && icmp_ping(addr))
		      {
			/* address in use: perturb address selection so that we are
			   less likely to try this address again. */
			if (!option_bool(OPT_CONSEC_ADDR))
			  c->addr_epoch++;
		      }
		    else
		      {
			/* at this point victim may hold an expired record */
			if (!victim)
			  {
			    if ((victim = whine_malloc(sizeof(struct ping_result))))
			      {
				victim->next = daemon->ping_results;
				daemon->ping_results = victim;
			      }
			  }
			
			/* record that this address is OK for 30s 
			   without more ping checks */
			if (victim)
			  {
			    victim->addr = addr;
			    victim->time = now;
			    victim->hash = j;
			  }
			return 1;
		      }
		  }
	      }

	    addr.s_addr = htonl(ntohl(addr.s_addr) + 1);
	    
	    if (addr.s_addr == htonl(ntohl(c->end.s_addr) + 1))
	      addr = c->start;
	    
	  } while (addr.s_addr != start.s_addr);
	}

  return 0;
}

void dhcp_read_ethers(void)
{
  FILE *f = fopen(ETHERSFILE, "r");
  unsigned int flags;
  char *buff = daemon->namebuff;
  char *ip, *cp;
  struct in_addr addr;
  unsigned char hwaddr[ETHER_ADDR_LEN];
  struct dhcp_config **up, *tmp;
  struct dhcp_config *config;
  int count = 0, lineno = 0;

  addr.s_addr = 0; /* eliminate warning */
  
  if (!f)
    {
      my_syslog(MS_DHCP | LOG_ERR, _("failed to read %s: %s"), ETHERSFILE, strerror(errno));
      return;
    }

  /* This can be called again on SIGHUP, so remove entries created last time round. */
  for (up = &daemon->dhcp_conf, config = daemon->dhcp_conf; config; config = tmp)
    {
      tmp = config->next;
      if (config->flags & CONFIG_FROM_ETHERS)
	{
	  *up = tmp;
	  /* cannot have a clid */
	  if (config->flags & CONFIG_NAME)
	    free(config->hostname);
	  free(config->hwaddr);
	  free(config);
	}
      else
	up = &config->next;
    }

  while (fgets(buff, MAXDNAME, f))
    {
      char *host = NULL;
      
      lineno++;
      
      while (strlen(buff) > 0 && isspace((int)buff[strlen(buff)-1]))
	buff[strlen(buff)-1] = 0;
      
      if ((*buff == '#') || (*buff == '+') || (*buff == 0))
	continue;
      
      for (ip = buff; *ip && !isspace((int)*ip); ip++);
      for(; *ip && isspace((int)*ip); ip++)
	*ip = 0;
      if (!*ip || parse_hex(buff, hwaddr, ETHER_ADDR_LEN, NULL, NULL) != ETHER_ADDR_LEN)
	{
	  my_syslog(MS_DHCP | LOG_ERR, _("bad line at %s line %d"), ETHERSFILE, lineno); 
	  continue;
	}
      
      /* check for name or dotted-quad */
      for (cp = ip; *cp; cp++)
	if (!(*cp == '.' || (*cp >='0' && *cp <= '9')))
	  break;
      
      if (!*cp)
	{
	  if ((addr.s_addr = inet_addr(ip)) == (in_addr_t)-1)
	    {
	      my_syslog(MS_DHCP | LOG_ERR, _("bad address at %s line %d"), ETHERSFILE, lineno); 
	      continue;
	    }

	  flags = CONFIG_ADDR;
	  
	  for (config = daemon->dhcp_conf; config; config = config->next)
	    if ((config->flags & CONFIG_ADDR) && config->addr.s_addr == addr.s_addr)
	      break;
	}
      else 
	{
	  int nomem;
	  if (!(host = canonicalise(ip, &nomem)) || !legal_hostname(host))
	    {
	      if (!nomem)
		my_syslog(MS_DHCP | LOG_ERR, _("bad name at %s line %d"), ETHERSFILE, lineno); 
	      free(host);
	      continue;
	    }
	      
	  flags = CONFIG_NAME;

	  for (config = daemon->dhcp_conf; config; config = config->next)
	    if ((config->flags & CONFIG_NAME) && hostname_isequal(config->hostname, host))
	      break;
	}

      if (config && (config->flags & CONFIG_FROM_ETHERS))
	{
	  my_syslog(MS_DHCP | LOG_ERR, _("ignoring %s line %d, duplicate name or IP address"), ETHERSFILE, lineno); 
	  continue;
	}
	
      if (!config)
	{ 
	  for (config = daemon->dhcp_conf; config; config = config->next)
	    {
	      struct hwaddr_config *conf_addr = config->hwaddr;
	      if (conf_addr && 
		  conf_addr->next == NULL && 
		  conf_addr->wildcard_mask == 0 &&
		  conf_addr->hwaddr_len == ETHER_ADDR_LEN &&
		  (conf_addr->hwaddr_type == ARPHRD_ETHER || conf_addr->hwaddr_type == 0) &&
		  memcmp(conf_addr->hwaddr, hwaddr, ETHER_ADDR_LEN) == 0)
		break;
	    }
	  
	  if (!config)
	    {
	      if (!(config = whine_malloc(sizeof(struct dhcp_config))))
		continue;
	      config->flags = CONFIG_FROM_ETHERS;
	      config->hwaddr = NULL;
	      config->domain = NULL;
	      config->netid = NULL;
	      config->next = daemon->dhcp_conf;
	      daemon->dhcp_conf = config;
	    }
	  
	  config->flags |= flags;
	  
	  if (flags & CONFIG_NAME)
	    {
	      config->hostname = host;
	      host = NULL;
	    }
	  
	  if (flags & CONFIG_ADDR)
	    config->addr = addr;
	}
      
      config->flags |= CONFIG_NOCLID;
      if (!config->hwaddr)
	config->hwaddr = whine_malloc(sizeof(struct hwaddr_config));
      if (config->hwaddr)
	{
	  memcpy(config->hwaddr->hwaddr, hwaddr, ETHER_ADDR_LEN);
	  config->hwaddr->hwaddr_len = ETHER_ADDR_LEN;
	  config->hwaddr->hwaddr_type = ARPHRD_ETHER;
	  config->hwaddr->wildcard_mask = 0;
	  config->hwaddr->next = NULL;
	}
      count++;
      
      free(host);

    }
  
  fclose(f);

  my_syslog(MS_DHCP | LOG_INFO, _("read %s - %d addresses"), ETHERSFILE, count);
}


/* If we've not found a hostname any other way, try and see if there's one in /etc/hosts
   for this address. If it has a domain part, that must match the set domain and
   it gets stripped. The set of legal domain names is bigger than the set of legal hostnames
   so check here that the domain name is legal as a hostname. 
   NOTE: we're only allowed to overwrite daemon->dhcp_buff if we succeed. */
char *host_from_dns(struct in_addr addr)
{
  struct crec *lookup;

  if (daemon->port == 0)
    return NULL; /* DNS disabled. */
  
  lookup = cache_find_by_addr(NULL, (struct all_addr *)&addr, 0, F_IPV4);

  if (lookup && (lookup->flags & F_HOSTS))
    {
      char *dot, *hostname = cache_get_name(lookup);
      dot = strchr(hostname, '.');
      
      if (dot && strlen(dot+1) != 0)
	{
	  char *d2 = get_domain(addr);
	  if (!d2 || !hostname_isequal(dot+1, d2))
	    return NULL; /* wrong domain */
	}

      if (!legal_hostname(hostname))
	return NULL;
      
      strncpy(daemon->dhcp_buff, hostname, 256);
      daemon->dhcp_buff[255] = 0;
      strip_hostname(daemon->dhcp_buff);

      return daemon->dhcp_buff;
    }
  
  return NULL;
}

static int  relay_upstream4(struct dhcp_relay *relay, struct dhcp_packet *mess, size_t sz, int iface_index)
{
  /* ->local is same value for all relays on ->current chain */
  struct all_addr from;
  
  if (mess->op != BOOTREQUEST)
    return 0;

  /* source address == relay address */
  from.addr.addr4 = relay->local.addr.addr4;
  
  /* already gatewayed ? */
  if (mess->giaddr.s_addr)
    {
      /* if so check if by us, to stomp on loops. */
      if (mess->giaddr.s_addr == relay->local.addr.addr4.s_addr)
	return 1;
    }
  else
    {
      /* plug in our address */
      mess->giaddr.s_addr = relay->local.addr.addr4.s_addr;
    }

  if ((mess->hops++) > 20)
    return 1;

  for (; relay; relay = relay->current)
    {
      union mysockaddr to;
      
      to.sa.sa_family = AF_INET;
      to.in.sin_addr = relay->server.addr.addr4;
      to.in.sin_port = htons(daemon->dhcp_server_port);
      
      send_from(daemon->dhcpfd, 0, (char *)mess, sz, &to, &from, 0);
      
      if (option_bool(OPT_LOG_OPTS))
	{
	  inet_ntop(AF_INET, &relay->local, daemon->addrbuff, ADDRSTRLEN);
	  my_syslog(MS_DHCP | LOG_INFO, _("DHCP relay %s -> %s"), daemon->addrbuff, inet_ntoa(relay->server.addr.addr4));
	}
      
      /* Save this for replies */
      relay->iface_index = iface_index;
    }
  
  return 1;
}


static struct dhcp_relay *relay_reply4(struct dhcp_packet *mess, char *arrival_interface)
{
  struct dhcp_relay *relay;

  if (mess->giaddr.s_addr == 0 || mess->op != BOOTREPLY)
    return NULL;

  for (relay = daemon->relay4; relay; relay = relay->next)
    {
      if (mess->giaddr.s_addr == relay->local.addr.addr4.s_addr)
	{
	  if (!relay->interface || wildcard_match(relay->interface, arrival_interface))
	    return relay->iface_index != 0 ? relay : NULL;
	}
    }
  
  return NULL;	 
}     

#endif