/*
       *	NET3	IP device support routines.
       *
       *	Version: $Id: devinet.c,v 1.39 2000/12/10 22:24:11 davem Exp $
       *
       *		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; either version
       *		2 of the License, or (at your option) any later version.
       *
       *	Derived from the IP parts of dev.c 1.0.19
       * 		Authors:	Ross Biro, <bir7@leland.Stanford.Edu>
       *				Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
       *				Mark Evans, <evansmp@uhura.aston.ac.uk>
       *
       *	Additional Authors:
       *		Alan Cox, <gw4pts@gw4pts.ampr.org>
       *		Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
       *
       *	Changes:
       *	        Alexey Kuznetsov:	pa_* fields are replaced with ifaddr lists.
       *		Cyrus Durgin:		updated for kmod
       */
      
      #include <linux/config.h>
       
      #include <asm/uaccess.h>
      #include <asm/system.h>
      #include <asm/bitops.h>
      #include <linux/types.h>
      #include <linux/kernel.h>
      #include <linux/sched.h>
      #include <linux/string.h>
      #include <linux/mm.h>
      #include <linux/socket.h>
      #include <linux/sockios.h>
      #include <linux/in.h>
      #include <linux/errno.h>
      #include <linux/interrupt.h>
      #include <linux/if_ether.h>
      #include <linux/inet.h>
      #include <linux/netdevice.h>
      #include <linux/etherdevice.h>
      #include <linux/skbuff.h>
      #include <linux/rtnetlink.h>
      #include <linux/init.h>
      #include <linux/notifier.h>
      #include <linux/inetdevice.h>
      #include <linux/igmp.h>
      #ifdef CONFIG_SYSCTL
      #include <linux/sysctl.h>
      #endif
      #ifdef CONFIG_KMOD
      #include <linux/kmod.h>
      #endif
      
      #include <net/ip.h>
      #include <net/route.h>
      #include <net/ip_fib.h>
      
      struct ipv4_devconf ipv4_devconf = { 1, 1, 1, 1, 0, };
      static struct ipv4_devconf ipv4_devconf_dflt = { 1, 1, 1, 1, 1, };
      
      #ifdef CONFIG_RTNETLINK
      static void rtmsg_ifa(int event, struct in_ifaddr *);
      #else
      #define rtmsg_ifa(a,b)	do { } while(0)
      #endif
      
      static struct notifier_block *inetaddr_chain;
      static void inet_del_ifa(struct in_device *in_dev, struct in_ifaddr **ifap, int destroy);
      #ifdef CONFIG_SYSCTL
      static void devinet_sysctl_register(struct in_device *in_dev, struct ipv4_devconf *p);
      static void devinet_sysctl_unregister(struct ipv4_devconf *p);
      #endif
      
      int inet_ifa_count;
      int inet_dev_count;
      
      /* Locks all the inet devices. */
      
      rwlock_t inetdev_lock = RW_LOCK_UNLOCKED;
      
      
  85  static struct in_ifaddr * inet_alloc_ifa(void)
      {
      	struct in_ifaddr *ifa;
      
      	ifa = kmalloc(sizeof(*ifa), GFP_KERNEL);
  90  	if (ifa) {
      		memset(ifa, 0, sizeof(*ifa));
      		inet_ifa_count++;
      	}
      
  95  	return ifa;
      }
      
  98  static __inline__ void inet_free_ifa(struct in_ifaddr *ifa)
      {
 100  	if (ifa->ifa_dev)
      		__in_dev_put(ifa->ifa_dev);
      	kfree(ifa);
      	inet_ifa_count--;
      }
      
 106  void in_dev_finish_destroy(struct in_device *idev)
      {
      	struct net_device *dev = idev->dev;
      
 110  	BUG_TRAP(idev->ifa_list==NULL);
 111  	BUG_TRAP(idev->mc_list==NULL);
      #ifdef NET_REFCNT_DEBUG
      	printk(KERN_DEBUG "in_dev_finish_destroy: %p=%s\n", idev, dev ? dev->name : "NIL");
      #endif
      	dev_put(dev);
 116  	if (!idev->dead) {
      		printk("Freeing alive in_device %p\n", idev);
 118  		return;
      	}
      	inet_dev_count--;
      	kfree(idev);
      }
      
 124  struct in_device *inetdev_init(struct net_device *dev)
      {
      	struct in_device *in_dev;
      
 128  	ASSERT_RTNL();
      
      	in_dev = kmalloc(sizeof(*in_dev), GFP_KERNEL);
 131  	if (!in_dev)
 132  		return NULL;
      	memset(in_dev, 0, sizeof(*in_dev));
      	in_dev->lock = RW_LOCK_UNLOCKED;
      	memcpy(&in_dev->cnf, &ipv4_devconf_dflt, sizeof(in_dev->cnf));
      	in_dev->cnf.sysctl = NULL;
      	in_dev->dev = dev;
 138  	if ((in_dev->arp_parms = neigh_parms_alloc(dev, &arp_tbl)) == NULL) {
      		kfree(in_dev);
 140  		return NULL;
      	}
      	inet_dev_count++;
      	/* Reference in_dev->dev */
      	dev_hold(dev);
      #ifdef CONFIG_SYSCTL
      	neigh_sysctl_register(dev, in_dev->arp_parms, NET_IPV4, NET_IPV4_NEIGH, "ipv4");
      #endif
 148  	write_lock_bh(&inetdev_lock);
      	dev->ip_ptr = in_dev;
      	/* Account for reference dev->ip_ptr */
      	in_dev_hold(in_dev);
 152  	write_unlock_bh(&inetdev_lock);
      #ifdef CONFIG_SYSCTL
      	devinet_sysctl_register(in_dev, &in_dev->cnf);
      #endif
 156  	if (dev->flags&IFF_UP)
      		ip_mc_up(in_dev);
 158  	return in_dev;
      }
      
 161  static void inetdev_destroy(struct in_device *in_dev)
      {
      	struct in_ifaddr *ifa;
      
 165  	ASSERT_RTNL();
      
      	in_dev->dead = 1;
      
      	ip_mc_destroy_dev(in_dev);
      
 171  	while ((ifa = in_dev->ifa_list) != NULL) {
      		inet_del_ifa(in_dev, &in_dev->ifa_list, 0);
      		inet_free_ifa(ifa);
      	}
      
      #ifdef CONFIG_SYSCTL
      	devinet_sysctl_unregister(&in_dev->cnf);
      #endif
 179  	write_lock_bh(&inetdev_lock);
      	in_dev->dev->ip_ptr = NULL;
      	/* in_dev_put following below will kill the in_device */
 182  	write_unlock_bh(&inetdev_lock);
      
      
      	neigh_parms_release(&arp_tbl, in_dev->arp_parms);
      	in_dev_put(in_dev);
      }
      
 189  int inet_addr_onlink(struct in_device *in_dev, u32 a, u32 b)
      {
      	read_lock(&in_dev->lock);
 192  	for_primary_ifa(in_dev) {
 193  		if (inet_ifa_match(a, ifa)) {
 194  			if (!b || inet_ifa_match(b, ifa)) {
 195  				read_unlock(&in_dev->lock);
 196  				return 1;
      			}
      		}
      	} endfor_ifa(in_dev);
 200  	read_unlock(&in_dev->lock);
 201  	return 0;
      } 
      
      static void
 205  inet_del_ifa(struct in_device *in_dev, struct in_ifaddr **ifap, int destroy)
      {
      	struct in_ifaddr *ifa1 = *ifap;
      
 209  	ASSERT_RTNL();
      
      	/* 1. Deleting primary ifaddr forces deletion all secondaries */
      
 213  	if (!(ifa1->ifa_flags&IFA_F_SECONDARY)) {
      		struct in_ifaddr *ifa;
      		struct in_ifaddr **ifap1 = &ifa1->ifa_next;
      
 217  		while ((ifa=*ifap1) != NULL) {
      			if (!(ifa->ifa_flags&IFA_F_SECONDARY) ||
      			    ifa1->ifa_mask != ifa->ifa_mask ||
 220  			    !inet_ifa_match(ifa1->ifa_address, ifa)) {
      				ifap1 = &ifa->ifa_next;
 222  				continue;
      			}
 224  			write_lock_bh(&in_dev->lock);
      			*ifap1 = ifa->ifa_next;
 226  			write_unlock_bh(&in_dev->lock);
      
 228  			rtmsg_ifa(RTM_DELADDR, ifa);
      			notifier_call_chain(&inetaddr_chain, NETDEV_DOWN, ifa);
      			inet_free_ifa(ifa);
      		}
      	}
      
      	/* 2. Unlink it */
      
 236  	write_lock_bh(&in_dev->lock);
      	*ifap = ifa1->ifa_next;
 238  	write_unlock_bh(&in_dev->lock);
      
      	/* 3. Announce address deletion */
      
      	/* Send message first, then call notifier.
      	   At first sight, FIB update triggered by notifier
      	   will refer to already deleted ifaddr, that could confuse
      	   netlink listeners. It is not true: look, gated sees
      	   that route deleted and if it still thinks that ifaddr
      	   is valid, it will try to restore deleted routes... Grr.
      	   So that, this order is correct.
      	 */
 250  	rtmsg_ifa(RTM_DELADDR, ifa1);
      	notifier_call_chain(&inetaddr_chain, NETDEV_DOWN, ifa1);
 252  	if (destroy) {
      		inet_free_ifa(ifa1);
      
 255  		if (in_dev->ifa_list == NULL)
      			inetdev_destroy(in_dev);
      	}
      }
      
      static int
 261  inet_insert_ifa(struct in_ifaddr *ifa)
      {
      	struct in_device *in_dev = ifa->ifa_dev;
      	struct in_ifaddr *ifa1, **ifap, **last_primary;
      
 266  	ASSERT_RTNL();
      
 268  	if (ifa->ifa_local == 0) {
      		inet_free_ifa(ifa);
 270  		return 0;
      	}
      
      	ifa->ifa_flags &= ~IFA_F_SECONDARY;
      	last_primary = &in_dev->ifa_list;
      
 276  	for (ifap=&in_dev->ifa_list; (ifa1=*ifap)!=NULL; ifap=&ifa1->ifa_next) {
 277  		if (!(ifa1->ifa_flags&IFA_F_SECONDARY) && ifa->ifa_scope <= ifa1->ifa_scope)
      			last_primary = &ifa1->ifa_next;
 279  		if (ifa1->ifa_mask == ifa->ifa_mask && inet_ifa_match(ifa1->ifa_address, ifa)) {
 280  			if (ifa1->ifa_local == ifa->ifa_local) {
      				inet_free_ifa(ifa);
 282  				return -EEXIST;
      			}
 284  			if (ifa1->ifa_scope != ifa->ifa_scope) {
      				inet_free_ifa(ifa);
 286  				return -EINVAL;
      			}
      			ifa->ifa_flags |= IFA_F_SECONDARY;
      		}
      	}
      
 292  	if (!(ifa->ifa_flags&IFA_F_SECONDARY)) {
      		net_srandom(ifa->ifa_local);
      		ifap = last_primary;
      	}
      
      	ifa->ifa_next = *ifap;
 298  	write_lock_bh(&in_dev->lock);
      	*ifap = ifa;
 300  	write_unlock_bh(&in_dev->lock);
      
      	/* Send message first, then call notifier.
      	   Notifier will trigger FIB update, so that
      	   listeners of netlink will know about new ifaddr */
 305  	rtmsg_ifa(RTM_NEWADDR, ifa);
      	notifier_call_chain(&inetaddr_chain, NETDEV_UP, ifa);
      
 308  	return 0;
      }
      
      static int
 312  inet_set_ifa(struct net_device *dev, struct in_ifaddr *ifa)
      {
      	struct in_device *in_dev = __in_dev_get(dev);
      
 316  	ASSERT_RTNL();
      
 318  	if (in_dev == NULL) {
      		in_dev = inetdev_init(dev);
 320  		if (in_dev == NULL) {
      			inet_free_ifa(ifa);
 322  			return -ENOBUFS;
      		}
      	}
 325  	if (ifa->ifa_dev != in_dev) {
 326  		BUG_TRAP(ifa->ifa_dev==NULL);
      		in_dev_hold(in_dev);
      		ifa->ifa_dev=in_dev;
      	}
 330  	if (LOOPBACK(ifa->ifa_local))
      		ifa->ifa_scope = RT_SCOPE_HOST;
 332  	return inet_insert_ifa(ifa);
      }
      
 335  struct in_device *inetdev_by_index(int ifindex)
      {
      	struct net_device *dev;
      	struct in_device *in_dev = NULL;
      	read_lock(&dev_base_lock);
      	dev = __dev_get_by_index(ifindex);
 341  	if (dev)
      		in_dev = in_dev_get(dev);
 343  	read_unlock(&dev_base_lock);
 344  	return in_dev;
      }
      
      /* Called only from RTNL semaphored context. No locks. */
      
 349  struct in_ifaddr *inet_ifa_byprefix(struct in_device *in_dev, u32 prefix, u32 mask)
      {
 351  	ASSERT_RTNL();
      
 353  	for_primary_ifa(in_dev) {
 354  		if (ifa->ifa_mask == mask && inet_ifa_match(prefix, ifa))
 355  			return ifa;
      	} endfor_ifa(in_dev);
 357  	return NULL;
      }
      
      #ifdef CONFIG_RTNETLINK
      
      int
      inet_rtm_deladdr(struct sk_buff *skb, struct nlmsghdr *nlh, void *arg)
      {
      	struct rtattr  **rta = arg;
      	struct in_device *in_dev;
      	struct ifaddrmsg *ifm = NLMSG_DATA(nlh);
      	struct in_ifaddr *ifa, **ifap;
      
      	ASSERT_RTNL();
      
      	if ((in_dev = inetdev_by_index(ifm->ifa_index)) == NULL)
      		return -EADDRNOTAVAIL;
      	__in_dev_put(in_dev);
      
      	for (ifap=&in_dev->ifa_list; (ifa=*ifap)!=NULL; ifap=&ifa->ifa_next) {
      		if ((rta[IFA_LOCAL-1] && memcmp(RTA_DATA(rta[IFA_LOCAL-1]), &ifa->ifa_local, 4)) ||
      		    (rta[IFA_LABEL-1] && strcmp(RTA_DATA(rta[IFA_LABEL-1]), ifa->ifa_label)) ||
      		    (rta[IFA_ADDRESS-1] &&
      		     (ifm->ifa_prefixlen != ifa->ifa_prefixlen ||
      		      !inet_ifa_match(*(u32*)RTA_DATA(rta[IFA_ADDRESS-1]), ifa))))
      			continue;
      		inet_del_ifa(in_dev, ifap, 1);
      		return 0;
      	}
      
      	return -EADDRNOTAVAIL;
      }
      
      int
      inet_rtm_newaddr(struct sk_buff *skb, struct nlmsghdr *nlh, void *arg)
      {
      	struct rtattr **rta = arg;
      	struct net_device *dev;
      	struct in_device *in_dev;
      	struct ifaddrmsg *ifm = NLMSG_DATA(nlh);
      	struct in_ifaddr *ifa;
      
      	ASSERT_RTNL();
      
      	if (ifm->ifa_prefixlen > 32 || rta[IFA_LOCAL-1] == NULL)
      		return -EINVAL;
      
      	if ((dev = __dev_get_by_index(ifm->ifa_index)) == NULL)
      		return -ENODEV;
      
      	if ((in_dev = __in_dev_get(dev)) == NULL) {
      		in_dev = inetdev_init(dev);
      		if (!in_dev)
      			return -ENOBUFS;
      	}
      
      	if ((ifa = inet_alloc_ifa()) == NULL)
      		return -ENOBUFS;
      
      	if (rta[IFA_ADDRESS-1] == NULL)
      		rta[IFA_ADDRESS-1] = rta[IFA_LOCAL-1];
      	memcpy(&ifa->ifa_local, RTA_DATA(rta[IFA_LOCAL-1]), 4);
      	memcpy(&ifa->ifa_address, RTA_DATA(rta[IFA_ADDRESS-1]), 4);
      	ifa->ifa_prefixlen = ifm->ifa_prefixlen;
      	ifa->ifa_mask = inet_make_mask(ifm->ifa_prefixlen);
      	if (rta[IFA_BROADCAST-1])
      		memcpy(&ifa->ifa_broadcast, RTA_DATA(rta[IFA_BROADCAST-1]), 4);
      	if (rta[IFA_ANYCAST-1])
      		memcpy(&ifa->ifa_anycast, RTA_DATA(rta[IFA_ANYCAST-1]), 4);
      	ifa->ifa_flags = ifm->ifa_flags;
      	ifa->ifa_scope = ifm->ifa_scope;
      	in_dev_hold(in_dev);
      	ifa->ifa_dev = in_dev;
      	if (rta[IFA_LABEL-1])
      		memcpy(ifa->ifa_label, RTA_DATA(rta[IFA_LABEL-1]), IFNAMSIZ);
      	else
      		memcpy(ifa->ifa_label, dev->name, IFNAMSIZ);
      
      	return inet_insert_ifa(ifa);
      }
      
      #endif
      
      /* 
       *	Determine a default network mask, based on the IP address. 
       */
      
 444  static __inline__ int inet_abc_len(u32 addr)
      {
 446    	if (ZERONET(addr))
 447    		return 0;
      
        	addr = ntohl(addr);
 450    	if (IN_CLASSA(addr)) 
 451    		return 8;
 452    	if (IN_CLASSB(addr)) 
 453    		return 16;
 454    	if (IN_CLASSC(addr)) 
 455    		return 24;
      
      	/*
      	 *	Something else, probably a multicast. 
      	 */
        	 
 461    	return -1;
      }
      
      
 465  int devinet_ioctl(unsigned int cmd, void *arg)
      {
      	struct ifreq ifr;
      	struct sockaddr_in *sin = (struct sockaddr_in *)&ifr.ifr_addr;
      	struct in_device *in_dev;
      	struct in_ifaddr **ifap = NULL;
      	struct in_ifaddr *ifa = NULL;
      	struct net_device *dev;
      	char *colon;
      	int ret = 0;
      
      	/*
      	 *	Fetch the caller's info block into kernel space
      	 */
      
 480  	if (copy_from_user(&ifr, arg, sizeof(struct ifreq)))
 481  		return -EFAULT;
      	ifr.ifr_name[IFNAMSIZ-1] = 0;
      
      	colon = strchr(ifr.ifr_name, ':');
 485  	if (colon)
      		*colon = 0;
      
      #ifdef CONFIG_KMOD
      	dev_load(ifr.ifr_name);
      #endif
      
 492  	switch(cmd) {
 493  	case SIOCGIFADDR:	/* Get interface address */
 494  	case SIOCGIFBRDADDR:	/* Get the broadcast address */
 495  	case SIOCGIFDSTADDR:	/* Get the destination address */
 496  	case SIOCGIFNETMASK:	/* Get the netmask for the interface */
      		/* Note that this ioctls will not sleep,
      		   so that we do not impose a lock.
      		   One day we will be forced to put shlock here (I mean SMP)
      		 */
      		memset(sin, 0, sizeof(*sin));
      		sin->sin_family = AF_INET;
 503  		break;
      
 505  	case SIOCSIFFLAGS:
 506  		if (!capable(CAP_NET_ADMIN))
 507  			return -EACCES;
 508  		break;
 509  	case SIOCSIFADDR:	/* Set interface address (and family) */
 510  	case SIOCSIFBRDADDR:	/* Set the broadcast address */
 511  	case SIOCSIFDSTADDR:	/* Set the destination address */
 512  	case SIOCSIFNETMASK: 	/* Set the netmask for the interface */
 513  		if (!capable(CAP_NET_ADMIN))
 514  			return -EACCES;
 515  		if (sin->sin_family != AF_INET)
 516  			return -EINVAL;
 517  		break;
 518  	default:
 519  		return -EINVAL;
      	}
      
      	dev_probe_lock();
      	rtnl_lock();
      
 525  	if ((dev = __dev_get_by_name(ifr.ifr_name)) == NULL) {
      		ret = -ENODEV;
 527  		goto done;
      	}
      
 530  	if (colon)
      		*colon = ':';
      
 533  	if ((in_dev=__in_dev_get(dev)) != NULL) {
 534  		for (ifap=&in_dev->ifa_list; (ifa=*ifap) != NULL; ifap=&ifa->ifa_next)
 535  			if (strcmp(ifr.ifr_name, ifa->ifa_label) == 0)
 536  				break;
      	}
      
 539  	if (ifa == NULL && cmd != SIOCSIFADDR && cmd != SIOCSIFFLAGS) {
      		ret = -EADDRNOTAVAIL;
 541  		goto done;
      	}
      
 544  	switch(cmd) {
 545  		case SIOCGIFADDR:	/* Get interface address */
      			sin->sin_addr.s_addr = ifa->ifa_local;
 547  			goto rarok;
      
 549  		case SIOCGIFBRDADDR:	/* Get the broadcast address */
      			sin->sin_addr.s_addr = ifa->ifa_broadcast;
 551  			goto rarok;
      
 553  		case SIOCGIFDSTADDR:	/* Get the destination address */
      			sin->sin_addr.s_addr = ifa->ifa_address;
 555  			goto rarok;
      
 557  		case SIOCGIFNETMASK:	/* Get the netmask for the interface */
      			sin->sin_addr.s_addr = ifa->ifa_mask;
 559  			goto rarok;
      
 561  		case SIOCSIFFLAGS:
 562  			if (colon) {
 563  				if (ifa == NULL) {
      					ret = -EADDRNOTAVAIL;
 565  					break;
      				}
 567  				if (!(ifr.ifr_flags&IFF_UP))
      					inet_del_ifa(in_dev, ifap, 1);
 569  				break;
      			}
      			ret = dev_change_flags(dev, ifr.ifr_flags);
 572  			break;
      	
 574  		case SIOCSIFADDR:	/* Set interface address (and family) */
 575  			if (inet_abc_len(sin->sin_addr.s_addr) < 0) {
      				ret = -EINVAL;
 577  				break;
      			}
      
 580  			if (!ifa) {
 581  				if ((ifa = inet_alloc_ifa()) == NULL) {
      					ret = -ENOBUFS;
 583  					break;
      				}
 585  				if (colon)
      					memcpy(ifa->ifa_label, ifr.ifr_name, IFNAMSIZ);
 587  				else
      					memcpy(ifa->ifa_label, dev->name, IFNAMSIZ);
 589  			} else {
      				ret = 0;
 591  				if (ifa->ifa_local == sin->sin_addr.s_addr)
 592  					break;
      				inet_del_ifa(in_dev, ifap, 0);
      				ifa->ifa_broadcast = 0;
      				ifa->ifa_anycast = 0;
      			}
      
      			ifa->ifa_address =
      			ifa->ifa_local = sin->sin_addr.s_addr;
      
 601  			if (!(dev->flags&IFF_POINTOPOINT)) {
      				ifa->ifa_prefixlen = inet_abc_len(ifa->ifa_address);
      				ifa->ifa_mask = inet_make_mask(ifa->ifa_prefixlen);
 604  				if ((dev->flags&IFF_BROADCAST) && ifa->ifa_prefixlen < 31)
      					ifa->ifa_broadcast = ifa->ifa_address|~ifa->ifa_mask;
 606  			} else {
      				ifa->ifa_prefixlen = 32;
      				ifa->ifa_mask = inet_make_mask(32);
      			}
      			ret = inet_set_ifa(dev, ifa);
 611  			break;
      
 613  		case SIOCSIFBRDADDR:	/* Set the broadcast address */
 614  			if (ifa->ifa_broadcast != sin->sin_addr.s_addr) {
      				inet_del_ifa(in_dev, ifap, 0);
      				ifa->ifa_broadcast = sin->sin_addr.s_addr;
      				inet_insert_ifa(ifa);
      			}
 619  			break;
      	
 621  		case SIOCSIFDSTADDR:	/* Set the destination address */
 622  			if (ifa->ifa_address != sin->sin_addr.s_addr) {
 623  				if (inet_abc_len(sin->sin_addr.s_addr) < 0) {
      					ret = -EINVAL;
 625  					break;
      				}
      				inet_del_ifa(in_dev, ifap, 0);
      				ifa->ifa_address = sin->sin_addr.s_addr;
      				inet_insert_ifa(ifa);
      			}
 631  			break;
      
 633  		case SIOCSIFNETMASK: 	/* Set the netmask for the interface */
      
      			/*
      			 *	The mask we set must be legal.
      			 */
 638  			if (bad_mask(sin->sin_addr.s_addr, 0)) {
      				ret = -EINVAL;
 640  				break;
      			}
      
 643  			if (ifa->ifa_mask != sin->sin_addr.s_addr) {
      				inet_del_ifa(in_dev, ifap, 0);
      				ifa->ifa_mask = sin->sin_addr.s_addr;
      				ifa->ifa_prefixlen = inet_mask_len(ifa->ifa_mask);
      				inet_insert_ifa(ifa);
      			}
 649  			break;
      	}
      done:
      	rtnl_unlock();
      	dev_probe_unlock();
 654  	return ret;
      
      rarok:
      	rtnl_unlock();
      	dev_probe_unlock();
 659  	if (copy_to_user(arg, &ifr, sizeof(struct ifreq)))
 660  		return -EFAULT;
 661  	return 0;
      }
      
      static int
 665  inet_gifconf(struct net_device *dev, char *buf, int len)
      {
      	struct in_device *in_dev = __in_dev_get(dev);
      	struct in_ifaddr *ifa;
      	struct ifreq ifr;
      	int done=0;
      
 672  	if (in_dev==NULL || (ifa=in_dev->ifa_list)==NULL)
 673  		return 0;
      
 675  	for ( ; ifa; ifa = ifa->ifa_next) {
 676  		if (!buf) {
      			done += sizeof(ifr);
 678  			continue;
      		}
 680  		if (len < (int) sizeof(ifr))
 681  			return done;
      		memset(&ifr, 0, sizeof(struct ifreq));
 683  		if (ifa->ifa_label)
      			strcpy(ifr.ifr_name, ifa->ifa_label);
 685  		else
      			strcpy(ifr.ifr_name, dev->name);
      
      		(*(struct sockaddr_in *) &ifr.ifr_addr).sin_family = AF_INET;
      		(*(struct sockaddr_in *) &ifr.ifr_addr).sin_addr.s_addr = ifa->ifa_local;
      
 691  		if (copy_to_user(buf, &ifr, sizeof(struct ifreq)))
 692  			return -EFAULT;
      		buf += sizeof(struct ifreq);
      		len -= sizeof(struct ifreq);
      		done += sizeof(struct ifreq);
      	}
 697  	return done;
      }
      
 700  u32 inet_select_addr(const struct net_device *dev, u32 dst, int scope)
      {
      	u32 addr = 0;
      	struct in_device *in_dev;
      
      	read_lock(&inetdev_lock);
      	in_dev = __in_dev_get(dev);
 707  	if (in_dev == NULL) {
 708  		read_unlock(&inetdev_lock);
 709  		return 0;
      	}
      
      	read_lock(&in_dev->lock);
 713  	for_primary_ifa(in_dev) {
 714  		if (ifa->ifa_scope > scope)
 715  			continue;
 716  		if (!dst || inet_ifa_match(dst, ifa)) {
      			addr = ifa->ifa_local;
 718  			break;
      		}
 720  		if (!addr)
      			addr = ifa->ifa_local;
      	} endfor_ifa(in_dev);
 723  	read_unlock(&in_dev->lock);
 724  	read_unlock(&inetdev_lock);
      
 726  	if (addr)
 727  		return addr;
      
      	/* Not loopback addresses on loopback should be preferred
      	   in this case. It is importnat that lo is the first interface
      	   in dev_base list.
      	 */
      	read_lock(&dev_base_lock);
      	read_lock(&inetdev_lock);
 735  	for (dev=dev_base; dev; dev=dev->next) {
 736  		if ((in_dev=__in_dev_get(dev)) == NULL)
 737  			continue;
      
      		read_lock(&in_dev->lock);
 740  		for_primary_ifa(in_dev) {
      			if (ifa->ifa_scope != RT_SCOPE_LINK &&
 742  			    ifa->ifa_scope <= scope) {
 743  				read_unlock(&in_dev->lock);
 744  				read_unlock(&inetdev_lock);
 745  				read_unlock(&dev_base_lock);
 746  				return ifa->ifa_local;
      			}
      		} endfor_ifa(in_dev);
 749  		read_unlock(&in_dev->lock);
      	}
 751  	read_unlock(&inetdev_lock);
 752  	read_unlock(&dev_base_lock);
      
 754  	return 0;
      }
      
      /*
       *	Device notifier
       */
      
 761  int register_inetaddr_notifier(struct notifier_block *nb)
      {
 763  	return notifier_chain_register(&inetaddr_chain, nb);
      }
      
 766  int unregister_inetaddr_notifier(struct notifier_block *nb)
      {
 768  	return notifier_chain_unregister(&inetaddr_chain,nb);
      }
      
      /* Called only under RTNL semaphore */
      
 773  static int inetdev_event(struct notifier_block *this, unsigned long event, void *ptr)
      {
      	struct net_device *dev = ptr;
      	struct in_device *in_dev = __in_dev_get(dev);
      
 778  	ASSERT_RTNL();
      
 780  	if (in_dev == NULL)
 781  		return NOTIFY_DONE;
      
 783  	switch (event) {
 784  	case NETDEV_REGISTER:
      		printk(KERN_DEBUG "inetdev_event: bug\n");
      		dev->ip_ptr = NULL;
 787  		break;
 788  	case NETDEV_UP:
 789  		if (dev->mtu < 68)
 790  			break;
 791  		if (dev == &loopback_dev) {
      			struct in_ifaddr *ifa;
 793  			if ((ifa = inet_alloc_ifa()) != NULL) {
      				ifa->ifa_local =
      				ifa->ifa_address = htonl(INADDR_LOOPBACK);
      				ifa->ifa_prefixlen = 8;
      				ifa->ifa_mask = inet_make_mask(8);
      				in_dev_hold(in_dev);
      				ifa->ifa_dev = in_dev;
      				ifa->ifa_scope = RT_SCOPE_HOST;
      				memcpy(ifa->ifa_label, dev->name, IFNAMSIZ);
      				inet_insert_ifa(ifa);
      			}
      		}
      		ip_mc_up(in_dev);
 806  		break;
 807  	case NETDEV_DOWN:
      		ip_mc_down(in_dev);
 809  		break;
 810  	case NETDEV_CHANGEMTU:
 811  		if (dev->mtu >= 68)
 812  			break;
      		/* MTU falled under 68, disable IP */
 814  	case NETDEV_UNREGISTER:
      		inetdev_destroy(in_dev);
 816  		break;
 817  	case NETDEV_CHANGENAME:
 818  		if (in_dev->ifa_list) {
      			struct in_ifaddr *ifa;
 820  			for (ifa = in_dev->ifa_list; ifa; ifa = ifa->ifa_next)
      				memcpy(ifa->ifa_label, dev->name, IFNAMSIZ);
      			/* Do not notify about label change, this event is
      			   not interesting to applications using netlink.
      			 */
      		}
 826  		break;
      	}
      
 829  	return NOTIFY_DONE;
      }
      
      struct notifier_block ip_netdev_notifier={
      	inetdev_event,
      	NULL,
      	0
      };
      
      #ifdef CONFIG_RTNETLINK
      
      static int inet_fill_ifaddr(struct sk_buff *skb, struct in_ifaddr *ifa,
      			    u32 pid, u32 seq, int event)
      {
      	struct ifaddrmsg *ifm;
      	struct nlmsghdr  *nlh;
      	unsigned char	 *b = skb->tail;
      
      	nlh = NLMSG_PUT(skb, pid, seq, event, sizeof(*ifm));
      	ifm = NLMSG_DATA(nlh);
      	ifm->ifa_family = AF_INET;
      	ifm->ifa_prefixlen = ifa->ifa_prefixlen;
      	ifm->ifa_flags = ifa->ifa_flags|IFA_F_PERMANENT;
      	ifm->ifa_scope = ifa->ifa_scope;
      	ifm->ifa_index = ifa->ifa_dev->dev->ifindex;
      	if (ifa->ifa_address)
      		RTA_PUT(skb, IFA_ADDRESS, 4, &ifa->ifa_address);
      	if (ifa->ifa_local)
      		RTA_PUT(skb, IFA_LOCAL, 4, &ifa->ifa_local);
      	if (ifa->ifa_broadcast)
      		RTA_PUT(skb, IFA_BROADCAST, 4, &ifa->ifa_broadcast);
      	if (ifa->ifa_anycast)
      		RTA_PUT(skb, IFA_ANYCAST, 4, &ifa->ifa_anycast);
      	if (ifa->ifa_label[0])
      		RTA_PUT(skb, IFA_LABEL, IFNAMSIZ, &ifa->ifa_label);
      	nlh->nlmsg_len = skb->tail - b;
      	return skb->len;
      
      nlmsg_failure:
      rtattr_failure:
      	skb_trim(skb, b - skb->data);
      	return -1;
      }
      
      static int inet_dump_ifaddr(struct sk_buff *skb, struct netlink_callback *cb)
      {
      	int idx, ip_idx;
      	int s_idx, s_ip_idx;
      	struct net_device *dev;
      	struct in_device *in_dev;
      	struct in_ifaddr *ifa;
      
      	s_idx = cb->args[0];
      	s_ip_idx = ip_idx = cb->args[1];
      	read_lock(&dev_base_lock);
      	for (dev=dev_base, idx=0; dev; dev = dev->next, idx++) {
      		if (idx < s_idx)
      			continue;
      		if (idx > s_idx)
      			s_ip_idx = 0;
      		read_lock(&inetdev_lock);
      		if ((in_dev = __in_dev_get(dev)) == NULL) {
      			read_unlock(&inetdev_lock);
      			continue;
      		}
      		read_lock(&in_dev->lock);
      		for (ifa = in_dev->ifa_list, ip_idx = 0; ifa;
      		     ifa = ifa->ifa_next, ip_idx++) {
      			if (ip_idx < s_ip_idx)
      				continue;
      			if (inet_fill_ifaddr(skb, ifa, NETLINK_CB(cb->skb).pid,
      					     cb->nlh->nlmsg_seq, RTM_NEWADDR) <= 0) {
      				read_unlock(&in_dev->lock);
      				read_unlock(&inetdev_lock);
      				goto done;
      			}
      		}
      		read_unlock(&in_dev->lock);
      		read_unlock(&inetdev_lock);
      	}
      
      done:
      	read_unlock(&dev_base_lock);
      	cb->args[0] = idx;
      	cb->args[1] = ip_idx;
      
      	return skb->len;
      }
      
      static void rtmsg_ifa(int event, struct in_ifaddr * ifa)
      {
      	struct sk_buff *skb;
      	int size = NLMSG_SPACE(sizeof(struct ifaddrmsg)+128);
      
      	skb = alloc_skb(size, GFP_KERNEL);
      	if (!skb) {
      		netlink_set_err(rtnl, 0, RTMGRP_IPV4_IFADDR, ENOBUFS);
      		return;
      	}
      	if (inet_fill_ifaddr(skb, ifa, 0, 0, event) < 0) {
      		kfree_skb(skb);
      		netlink_set_err(rtnl, 0, RTMGRP_IPV4_IFADDR, EINVAL);
      		return;
      	}
      	NETLINK_CB(skb).dst_groups = RTMGRP_IPV4_IFADDR;
      	netlink_broadcast(rtnl, skb, 0, RTMGRP_IPV4_IFADDR, GFP_KERNEL);
      }
      
      
      static struct rtnetlink_link inet_rtnetlink_table[RTM_MAX-RTM_BASE+1] =
      {
      	{ NULL,			NULL,			},
      	{ NULL,			NULL,			},
      	{ NULL,			NULL,			},
      	{ NULL,			NULL,			},
      
      	{ inet_rtm_newaddr,	NULL,			},
      	{ inet_rtm_deladdr,	NULL,			},
      	{ NULL,			inet_dump_ifaddr,	},
      	{ NULL,			NULL,			},
      
      	{ inet_rtm_newroute,	NULL,			},
      	{ inet_rtm_delroute,	NULL,			},
      	{ inet_rtm_getroute,	inet_dump_fib,		},
      	{ NULL,			NULL,			},
      
      	{ NULL,			NULL,			},
      	{ NULL,			NULL,			},
      	{ NULL,			NULL,			},
      	{ NULL,			NULL,			},
      
      #ifdef CONFIG_IP_MULTIPLE_TABLES
      	{ inet_rtm_newrule,	NULL,			},
      	{ inet_rtm_delrule,	NULL,			},
      	{ NULL,			inet_dump_rules,	},
      	{ NULL,			NULL,			},
      #else
      	{ NULL,			NULL,			},
      	{ NULL,			NULL,			},
      	{ NULL,			NULL,			},
      	{ NULL,			NULL,			},
      #endif
      };
      
      #endif /* CONFIG_RTNETLINK */
      
      
      #ifdef CONFIG_SYSCTL
      
 978  void inet_forward_change()
      {
      	struct net_device *dev;
      	int on = ipv4_devconf.forwarding;
      
      	ipv4_devconf.accept_redirects = !on;
      	ipv4_devconf_dflt.forwarding = on;
      
      	read_lock(&dev_base_lock);
 987  	for (dev = dev_base; dev; dev = dev->next) {
      		struct in_device *in_dev;
      		read_lock(&inetdev_lock);
      		in_dev = __in_dev_get(dev);
 991  		if (in_dev)
      			in_dev->cnf.forwarding = on;
 993  		read_unlock(&inetdev_lock);
      	}
 995  	read_unlock(&dev_base_lock);
      
      	rt_cache_flush(0);
      }
      
      static
1001  int devinet_sysctl_forward(ctl_table *ctl, int write, struct file * filp,
      			   void *buffer, size_t *lenp)
      {
      	int *valp = ctl->data;
      	int val = *valp;
      	int ret;
      
      	ret = proc_dointvec(ctl, write, filp, buffer, lenp);
      
1010  	if (write && *valp != val) {
1011  		if (valp == &ipv4_devconf.forwarding)
      			inet_forward_change();
1013  		else if (valp != &ipv4_devconf_dflt.forwarding)
      			rt_cache_flush(0);
      	}
      
1017          return ret;
      }
      
      static struct devinet_sysctl_table
      {
      	struct ctl_table_header *sysctl_header;
      	ctl_table devinet_vars[13];
      	ctl_table devinet_dev[2];
      	ctl_table devinet_conf_dir[2];
      	ctl_table devinet_proto_dir[2];
      	ctl_table devinet_root_dir[2];
      } devinet_sysctl = {
      	NULL,
      	{{NET_IPV4_CONF_FORWARDING, "forwarding",
               &ipv4_devconf.forwarding, sizeof(int), 0644, NULL,
               &devinet_sysctl_forward},
      	{NET_IPV4_CONF_MC_FORWARDING, "mc_forwarding",
               &ipv4_devconf.mc_forwarding, sizeof(int), 0444, NULL,
               &proc_dointvec},
      	{NET_IPV4_CONF_ACCEPT_REDIRECTS, "accept_redirects",
               &ipv4_devconf.accept_redirects, sizeof(int), 0644, NULL,
               &proc_dointvec},
      	{NET_IPV4_CONF_SECURE_REDIRECTS, "secure_redirects",
               &ipv4_devconf.secure_redirects, sizeof(int), 0644, NULL,
               &proc_dointvec},
      	{NET_IPV4_CONF_SHARED_MEDIA, "shared_media",
               &ipv4_devconf.shared_media, sizeof(int), 0644, NULL,
               &proc_dointvec},
      	{NET_IPV4_CONF_RP_FILTER, "rp_filter",
               &ipv4_devconf.rp_filter, sizeof(int), 0644, NULL,
               &proc_dointvec},
      	{NET_IPV4_CONF_SEND_REDIRECTS, "send_redirects",
               &ipv4_devconf.send_redirects, sizeof(int), 0644, NULL,
               &proc_dointvec},
      	{NET_IPV4_CONF_ACCEPT_SOURCE_ROUTE, "accept_source_route",
               &ipv4_devconf.accept_source_route, sizeof(int), 0644, NULL,
               &proc_dointvec},
      	{NET_IPV4_CONF_PROXY_ARP, "proxy_arp",
               &ipv4_devconf.proxy_arp, sizeof(int), 0644, NULL,
               &proc_dointvec},
      	{NET_IPV4_CONF_BOOTP_RELAY, "bootp_relay",
               &ipv4_devconf.bootp_relay, sizeof(int), 0644, NULL,
               &proc_dointvec},
              {NET_IPV4_CONF_LOG_MARTIANS, "log_martians",
               &ipv4_devconf.log_martians, sizeof(int), 0644, NULL,
               &proc_dointvec},
      	{NET_IPV4_CONF_TAG, "tag",
      	 &ipv4_devconf.tag, sizeof(int), 0644, NULL,
      	 &proc_dointvec},
      	 {0}},
      
      	{{NET_PROTO_CONF_ALL, "all", NULL, 0, 0555, devinet_sysctl.devinet_vars},{0}},
      	{{NET_IPV4_CONF, "conf", NULL, 0, 0555, devinet_sysctl.devinet_dev},{0}},
      	{{NET_IPV4, "ipv4", NULL, 0, 0555, devinet_sysctl.devinet_conf_dir},{0}},
      	{{CTL_NET, "net", NULL, 0, 0555, devinet_sysctl.devinet_proto_dir},{0}}
      };
      
1074  static void devinet_sysctl_register(struct in_device *in_dev, struct ipv4_devconf *p)
      {
      	int i;
      	struct net_device *dev = in_dev ? in_dev->dev : NULL;
      	struct devinet_sysctl_table *t;
      
      	t = kmalloc(sizeof(*t), GFP_KERNEL);
1081  	if (t == NULL)
1082  		return;
      	memcpy(t, &devinet_sysctl, sizeof(*t));
1084  	for (i=0; i<sizeof(t->devinet_vars)/sizeof(t->devinet_vars[0])-1; i++) {
      		t->devinet_vars[i].data += (char*)p - (char*)&ipv4_devconf;
      		t->devinet_vars[i].de = NULL;
      	}
1088  	if (dev) {
      		t->devinet_dev[0].procname = dev->name;
      		t->devinet_dev[0].ctl_name = dev->ifindex;
1091  	} else {
      		t->devinet_dev[0].procname = "default";
      		t->devinet_dev[0].ctl_name = NET_PROTO_CONF_DEFAULT;
      	}
      	t->devinet_dev[0].child = t->devinet_vars;
      	t->devinet_dev[0].de = NULL;
      	t->devinet_conf_dir[0].child = t->devinet_dev;
      	t->devinet_conf_dir[0].de = NULL;
      	t->devinet_proto_dir[0].child = t->devinet_conf_dir;
      	t->devinet_proto_dir[0].de = NULL;
      	t->devinet_root_dir[0].child = t->devinet_proto_dir;
      	t->devinet_root_dir[0].de = NULL;
      
      	t->sysctl_header = register_sysctl_table(t->devinet_root_dir, 0);
1105  	if (t->sysctl_header == NULL)
      		kfree(t);
1107  	else
      		p->sysctl = t;
      }
      
1111  static void devinet_sysctl_unregister(struct ipv4_devconf *p)
      {
1113  	if (p->sysctl) {
      		struct devinet_sysctl_table *t = p->sysctl;
      		p->sysctl = NULL;
      		unregister_sysctl_table(t->sysctl_header);
      		kfree(t);
      	}
      }
      #endif
      
1122  void __init devinet_init(void)
      {
      	register_gifconf(PF_INET, inet_gifconf);
      	register_netdevice_notifier(&ip_netdev_notifier);
      #ifdef CONFIG_RTNETLINK
      	rtnetlink_links[PF_INET] = inet_rtnetlink_table;
      #endif
      #ifdef CONFIG_SYSCTL
      	devinet_sysctl.sysctl_header =
      		register_sysctl_table(devinet_sysctl.devinet_root_dir, 0);
      	devinet_sysctl_register(NULL, &ipv4_devconf_dflt);
      #endif
      }