Merge with Linux 2.1.72, part 2.

The new signal code with exception of the code for the rt signals.
The definitions in <asm/siginfo.h> and <asm/ucontext.h> are currently
just stolen from the Alpha and will need to be overhauled.

9 files changed, 4948 insertions, 0 deletions

diff --git a/net/ipv4/fib.c b/net/ipv4/fib.c
deleted file mode 100644
index e69de29bb..000000000
--- a/net/ipv4/fib.c
+++ /dev/null
diff --git a/net/ipv4/fib_frontend.c b/net/ipv4/fib_frontend.c
new file mode 100644
index 000000000..16d72fcd2
--- /dev/null
+++ b/net/ipv4/fib_frontend.c
@@ -0,0 +1,572 @@
+/*
+ * INET		An implementation of the TCP/IP protocol suite for the LINUX
+ *		operating system.  INET is implemented using the  BSD Socket
+ *		interface as the means of communication with the user level.
+ *
+ *		IPv4 Forwarding Information Base: FIB frontend.
+ *
+ * Version:	$Id: fib_frontend.c,v 1.4 1997/11/09 20:05:23 kuznet Exp $
+ *
+ * Authors:	Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
+ *
+ *		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.
+ */
+
+#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/mm.h>
+#include <linux/string.h>
+#include <linux/socket.h>
+#include <linux/sockios.h>
+#include <linux/errno.h>
+#include <linux/in.h>
+#include <linux/inet.h>
+#include <linux/netdevice.h>
+#include <linux/if_arp.h>
+#include <linux/proc_fs.h>
+#include <linux/skbuff.h>
+#include <linux/netlink.h>
+#include <linux/init.h>
+
+#include <net/ip.h>
+#include <net/protocol.h>
+#include <net/route.h>
+#include <net/tcp.h>
+#include <net/sock.h>
+#include <net/icmp.h>
+#include <net/arp.h>
+#include <net/ip_fib.h>
+
+#define FFprint(a...) printk(KERN_DEBUG a)
+
+#ifndef CONFIG_IP_MULTIPLE_TABLES
+
+#define RT_TABLE_MIN RT_TABLE_MAIN
+
+struct fib_table *local_table;
+struct fib_table *main_table;
+
+#else
+
+#define RT_TABLE_MIN 1
+
+struct fib_table *fib_tables[RT_TABLE_MAX+1];
+
+struct fib_table *__fib_new_table(int id)
+{
+	struct fib_table *tb;
+
+	tb = fib_hash_init(id);
+	if (!tb)
+		return NULL;
+	fib_tables[id] = tb;
+	return tb;
+}
+
+
+#endif /* CONFIG_IP_MULTIPLE_TABLES */
+
+
+void fib_flush(void)
+{
+	int flushed = 0;
+#ifdef CONFIG_IP_MULTIPLE_TABLES
+	struct fib_table *tb;
+	int id;
+
+	for (id = RT_TABLE_MAX; id>0; id--) {
+		if ((tb = fib_get_table(id))==NULL)
+			continue;
+		flushed += tb->tb_flush(tb);
+	}
+#else /* CONFIG_IP_MULTIPLE_TABLES */
+	flushed += main_table->tb_flush(main_table);
+	flushed += local_table->tb_flush(local_table);
+#endif /* CONFIG_IP_MULTIPLE_TABLES */
+
+	if (flushed)
+		rt_cache_flush(RT_FLUSH_DELAY);
+}
+
+
+#ifdef CONFIG_PROC_FS
+
+/* 
+ *	Called from the PROCfs module. This outputs /proc/net/route.
+ *
+ *	It always works in backward compatibility mode.
+ *	The format of the file is not supposed to be changed.
+ */
+ 
+static int
+fib_get_procinfo(char *buffer, char **start, off_t offset, int length, int dummy)
+{
+	int first = offset/128;
+	char *ptr = buffer;
+	int count = (length+127)/128;
+	int len;
+
+	*start = buffer + offset%128;
+	
+	if (--first < 0) {
+		sprintf(buffer, "%-127s\n", "Iface\tDestination\tGateway \tFlags\tRefCnt\tUse\tMetric\tMask\t\tMTU\tWindow\tIRTT");
+		--count;
+		ptr += 128;
+		first = 0;
+  	}
+
+	/* rtnl_shlock(); -- it is pointless at the moment --ANK */
+	if (main_table && count > 0) {
+		int n = main_table->tb_get_info(main_table, ptr, first, count);
+		count -= n;
+		ptr += n*128;
+	}
+	/* rtnl_shunlock(); */
+	len = ptr - *start;
+	if (len >= length)
+		return length;
+	if (len >= 0)
+		return len;
+	return 0;
+}
+
+#endif /* CONFIG_PROC_FS */
+
+/*
+ *	Find the first device with a given source address.
+ */
+
+struct device * ip_dev_find(u32 addr)
+{
+	struct rt_key key;
+	struct fib_result res;
+
+	memset(&key, 0, sizeof(key));
+	key.dst = addr;
+	key.scope = RT_SCOPE_UNIVERSE;
+
+	if (!local_table || local_table->tb_lookup(local_table, &key, &res)
+	    || res.type != RTN_LOCAL)
+		return NULL;
+
+	return FIB_RES_DEV(res);
+}
+
+unsigned inet_addr_type(u32 addr)
+{
+	struct rt_key		key;
+	struct fib_result	res;
+
+	if (ZERONET(addr) || BADCLASS(addr))
+		return RTN_BROADCAST;
+	if (MULTICAST(addr))
+		return RTN_MULTICAST;
+
+	memset(&key, 0, sizeof(key));
+	key.dst = addr;
+
+	if (local_table) {
+		if (local_table->tb_lookup(local_table, &key, &res) == 0)
+			return res.type;
+		return RTN_UNICAST;
+	}
+	return RTN_BROADCAST;
+}
+
+/* Given (packet source, input interface) and optional (dst, oif, tos):
+   - (main) check, that source is valid i.e. not broadcast or our local
+     address.
+   - figure out what "logical" interface this packet arrived
+     and calculate "specific destination" address.
+   - check, that packet arrived from expected physical interface.
+ */
+
+int fib_validate_source(u32 src, u32 dst, u8 tos, int oif,
+			struct device *dev, u32 *spec_dst)
+{
+	struct in_device *in_dev = dev->ip_ptr;
+	struct rt_key key;
+	struct fib_result res;
+
+	key.dst = src;
+	key.src = dst;
+	key.tos = tos;
+	key.oif = 0;
+	key.iif = oif;
+	key.scope = RT_SCOPE_UNIVERSE;
+
+	if (in_dev == NULL)
+		return -EINVAL;
+	if (fib_lookup(&key, &res))
+		goto last_resort;
+	if (res.type != RTN_UNICAST)
+		return -EINVAL;
+	*spec_dst = FIB_RES_PREFSRC(res);
+#ifdef CONFIG_IP_ROUTE_MULTIPATH
+	if (FIB_RES_DEV(res) == dev || res.fi->fib_nhs > 1)
+#else
+	if (FIB_RES_DEV(res) == dev)
+#endif
+		return FIB_RES_NH(res).nh_scope >= RT_SCOPE_HOST;
+
+	if (in_dev->ifa_list == NULL)
+		goto last_resort;
+	if (IN_DEV_RPFILTER(in_dev))
+		return -EINVAL;
+	key.oif = dev->ifindex;
+	if (fib_lookup(&key, &res) == 0 && res.type == RTN_UNICAST) {
+		*spec_dst = FIB_RES_PREFSRC(res);
+		return FIB_RES_NH(res).nh_scope >= RT_SCOPE_HOST;
+	}
+	return 0;
+
+last_resort:
+	if (IN_DEV_RPFILTER(in_dev))
+		return -EINVAL;
+	*spec_dst = inet_select_addr(dev, 0, RT_SCOPE_UNIVERSE);
+	return 0;
+}
+
+#ifndef CONFIG_IP_NOSIOCRT
+
+/*
+ *	Handle IP routing ioctl calls. These are used to manipulate the routing tables
+ */
+ 
+int ip_rt_ioctl(unsigned int cmd, void *arg)
+{
+	int err;
+	struct kern_rta rta;
+	struct rtentry  r;
+	struct {
+		struct nlmsghdr nlh;
+		struct rtmsg	rtm;
+	} req;
+
+	switch (cmd) {
+	case SIOCADDRT:		/* Add a route */
+	case SIOCDELRT:		/* Delete a route */
+		if (!suser())
+			return -EPERM;
+		if (copy_from_user(&r, arg, sizeof(struct rtentry)))
+			return -EFAULT;
+		rtnl_lock();
+		err = fib_convert_rtentry(cmd, &req.nlh, &req.rtm, &rta, arg);
+		if (err == 0) {
+			if (cmd == SIOCDELRT) {
+				struct fib_table *tb = fib_get_table(req.rtm.rtm_table);
+				err = -ESRCH;
+				if (tb)
+					err = tb->tb_delete(tb, &req.rtm, &rta, &req.nlh, NULL);
+			} else {
+				struct fib_table *tb = fib_new_table(req.rtm.rtm_table);
+				err = -ENOBUFS;
+				if (tb)
+					err = tb->tb_insert(tb, &req.rtm, &rta, &req.nlh, NULL);
+			}
+		}
+		rtnl_unlock();
+		return err;
+	}
+	return -EINVAL;
+}
+
+#else
+
+int ip_rt_ioctl(unsigned int cmd, void *arg)
+{
+	return -EINVAL;
+}
+
+#endif
+
+#ifdef CONFIG_RTNETLINK
+
+int inet_rtm_delroute(struct sk_buff *skb, struct nlmsghdr* nlh, void *arg)
+{
+	struct fib_table * tb;
+	struct kern_rta *rta = arg;
+	struct rtmsg *r = NLMSG_DATA(nlh);
+
+	tb = fib_get_table(r->rtm_table);
+	if (tb)
+		return tb->tb_delete(tb, r, rta, nlh, &NETLINK_CB(skb));
+	return -ESRCH;
+}
+
+int inet_rtm_newroute(struct sk_buff *skb, struct nlmsghdr* nlh, void *arg)
+{
+	struct fib_table * tb;
+	struct kern_rta *rta = arg;
+	struct rtmsg *r = NLMSG_DATA(nlh);
+
+	tb = fib_new_table(r->rtm_table);
+	if (tb)
+		return tb->tb_insert(tb, r, rta, nlh, &NETLINK_CB(skb));
+	return -ENOBUFS;
+}
+
+int inet_dump_fib(struct sk_buff *skb, struct netlink_callback *cb)
+{
+	int t;
+	int s_t;
+	struct fib_table *tb;
+
+	s_t = cb->args[0];
+	if (s_t == 0)
+		s_t = cb->args[0] = RT_TABLE_MIN;
+
+	for (t=s_t; t<=RT_TABLE_MAX; t++) {
+		if (t < s_t) continue;
+		if (t > s_t)
+			memset(&cb->args[1], 0, sizeof(cb->args)-sizeof(int));
+		if ((tb = fib_get_table(t))==NULL)
+			continue;
+		if (tb->tb_dump(tb, skb, cb) < 0) 
+			break;
+	}
+
+	cb->args[0] = t;
+
+	return skb->len;
+}
+
+#endif
+
+/* Prepare and feed intra-kernel routing request.
+   Really, it should be netlink message, but :-( netlink
+   can be not configured, so that we feed it directly
+   to fib engine. It is legal, because all events occur
+   only when netlink is already locked.
+ */
+
+static void fib_magic(int cmd, int type, u32 dst, int dst_len, struct in_ifaddr *ifa)
+{
+	struct fib_table * tb;
+	struct {
+		struct nlmsghdr	nlh;
+		struct rtmsg	rtm;
+	} req;
+	struct kern_rta rta;
+
+	memset(&req.rtm, 0, sizeof(req.rtm));
+	memset(&rta, 0, sizeof(rta));
+
+	if (type == RTN_UNICAST)
+		tb = fib_new_table(RT_TABLE_MAIN);
+	else
+		tb = fib_new_table(RT_TABLE_LOCAL);
+
+	if (tb == NULL)
+		return;
+
+	req.nlh.nlmsg_len = sizeof(req);
+	req.nlh.nlmsg_type = cmd;
+	req.nlh.nlmsg_flags = NLM_F_REQUEST|NLM_F_CREATE;
+	req.nlh.nlmsg_pid = 0;
+	req.nlh.nlmsg_seq = 0;
+
+	req.rtm.rtm_dst_len = dst_len;
+	req.rtm.rtm_table = tb->tb_id;
+	req.rtm.rtm_protocol = RTPROT_KERNEL;
+	req.rtm.rtm_scope = (type != RTN_LOCAL ? RT_SCOPE_LINK : RT_SCOPE_HOST);
+	req.rtm.rtm_type = type;
+
+	rta.rta_dst = &dst;
+	rta.rta_prefsrc = &ifa->ifa_local;
+	rta.rta_oif = &ifa->ifa_dev->dev->ifindex;
+
+	if (cmd == RTM_NEWROUTE)
+		tb->tb_insert(tb, &req.rtm, &rta, &req.nlh, NULL);
+	else
+		tb->tb_delete(tb, &req.rtm, &rta, &req.nlh, NULL);
+}
+
+static void fib_add_ifaddr(struct in_ifaddr *ifa)
+{
+	struct in_device *in_dev = ifa->ifa_dev;
+	struct device *dev = in_dev->dev;
+	struct in_ifaddr *prim = ifa;
+	u32 mask = ifa->ifa_mask;
+	u32 addr = ifa->ifa_local;
+	u32 prefix = ifa->ifa_address&mask;
+
+	if (ifa->ifa_flags&IFA_F_SECONDARY)
+		prim = inet_ifa_byprefix(in_dev, prefix, mask);
+
+	fib_magic(RTM_NEWROUTE, RTN_LOCAL, addr, 32, prim);
+
+	if (!(dev->flags&IFF_UP))
+		return;
+
+	/* Add broadcast address, if it is explicitly assigned. */
+	if (ifa->ifa_broadcast && ifa->ifa_broadcast != 0xFFFFFFFF)
+		fib_magic(RTM_NEWROUTE, RTN_BROADCAST, ifa->ifa_broadcast, 32, prim);
+
+	if (!ZERONET(prefix) && !(ifa->ifa_flags&IFA_F_SECONDARY)) {
+		fib_magic(RTM_NEWROUTE, dev->flags&IFF_LOOPBACK ? RTN_LOCAL :
+			  RTN_UNICAST, prefix, ifa->ifa_prefixlen, prim);
+
+		/* Add network specific broadcasts, when it takes a sense */
+		if (ifa->ifa_prefixlen < 31) {
+			fib_magic(RTM_NEWROUTE, RTN_BROADCAST, prefix, 32, prim);
+			fib_magic(RTM_NEWROUTE, RTN_BROADCAST, prefix|~mask, 32, prim);
+		}
+	}
+}
+
+static void fib_del_ifaddr(struct in_ifaddr *ifa)
+{
+	struct in_device *in_dev = ifa->ifa_dev;
+	struct device *dev = in_dev->dev;
+	struct in_ifaddr *ifa1;
+	struct in_ifaddr *prim = ifa;
+	u32 brd = ifa->ifa_address|~ifa->ifa_mask;
+	u32 any = ifa->ifa_address&ifa->ifa_mask;
+#define LOCAL_OK	1
+#define BRD_OK		2
+#define BRD0_OK		4
+#define BRD1_OK		8
+	unsigned ok = 0;
+
+	if (!(ifa->ifa_flags&IFA_F_SECONDARY))
+		fib_magic(RTM_DELROUTE, dev->flags&IFF_LOOPBACK ? RTN_LOCAL :
+			  RTN_UNICAST, any, ifa->ifa_prefixlen, prim);
+	else 
+		prim = inet_ifa_byprefix(in_dev, any, ifa->ifa_mask);
+
+	/* Deletion is more complicated than add.
+	   We should take care of not to delete too much :-)
+
+	   Scan address list to be sure that addresses are really gone.
+	 */
+
+	for (ifa1 = in_dev->ifa_list; ifa1; ifa1 = ifa1->ifa_next) {
+		if (ifa->ifa_local == ifa1->ifa_local)
+			ok |= LOCAL_OK;
+		if (ifa->ifa_broadcast == ifa1->ifa_broadcast)
+			ok |= BRD_OK;
+		if (brd == ifa1->ifa_broadcast)
+			ok |= BRD1_OK;
+		if (any == ifa1->ifa_broadcast)
+			ok |= BRD0_OK;
+	}
+
+	if (!(ok&BRD_OK))
+		fib_magic(RTM_DELROUTE, RTN_BROADCAST, ifa->ifa_broadcast, 32, prim);
+	if (!(ok&BRD1_OK))
+		fib_magic(RTM_DELROUTE, RTN_BROADCAST, brd, 32, prim);
+	if (!(ok&BRD0_OK))
+		fib_magic(RTM_DELROUTE, RTN_BROADCAST, any, 32, prim);
+	if (!(ok&LOCAL_OK)) {
+		fib_magic(RTM_DELROUTE, RTN_LOCAL, ifa->ifa_local, 32, prim);
+
+		/* Check, that this local address finally disappeared. */
+		if (inet_addr_type(ifa->ifa_local) != RTN_LOCAL) {
+			/* And the last, but not the least thing.
+			   We must flush stray FIB entries.
+
+			   First of all, we scan fib_info list searching
+			   for stray nexthop entries, then ignite fib_flush.
+			*/
+			if (fib_sync_down(ifa->ifa_local, NULL))
+				fib_flush();
+		}
+	}
+#undef LOCAL_OK
+#undef BRD_OK
+#undef BRD0_OK
+#undef BRD1_OK
+}
+
+static int fib_inetaddr_event(struct notifier_block *this, unsigned long event, void *ptr)
+{
+	struct in_ifaddr *ifa = (struct in_ifaddr*)ptr;
+
+	switch (event) {
+	case NETDEV_UP:
+		fib_add_ifaddr(ifa);
+		rt_cache_flush(2*HZ);
+		break;
+	case NETDEV_DOWN:
+		fib_del_ifaddr(ifa);
+		rt_cache_flush(1*HZ);
+		break;
+	}
+	return NOTIFY_DONE;
+}
+
+static int fib_netdev_event(struct notifier_block *this, unsigned long event, void *ptr)
+{
+	struct device *dev = ptr;
+	struct in_device *in_dev = dev->ip_ptr;
+
+	if (!in_dev)
+		return NOTIFY_DONE;
+
+	switch (event) {
+	case NETDEV_UP:
+		for_ifa(in_dev) {
+			fib_add_ifaddr(ifa);
+		} endfor_ifa(in_dev);
+#ifdef CONFIG_IP_ROUTE_MULTIPATH
+		fib_sync_up(dev);
+#endif
+		rt_cache_flush(2*HZ);
+		break;
+	case NETDEV_DOWN:
+		if (fib_sync_down(0, dev))
+			fib_flush();
+		rt_cache_flush(0);
+		break;
+	case NETDEV_UNREGISTER:
+		if (in_dev->ifa_list)
+			printk("About to crash!\n");
+		rt_cache_flush(0);
+		break;
+	}
+	return NOTIFY_DONE;
+}
+
+struct notifier_block fib_inetaddr_notifier = {
+	fib_inetaddr_event,
+	NULL,
+	0
+};
+
+struct notifier_block fib_netdev_notifier = {
+	fib_netdev_event,
+	NULL,
+	0
+};
+
+__initfunc(void ip_fib_init(void))
+{
+#ifdef CONFIG_PROC_FS
+	proc_net_register(&(struct proc_dir_entry) {
+		PROC_NET_ROUTE, 5, "route",
+		S_IFREG | S_IRUGO, 1, 0, 0,
+		0, &proc_net_inode_operations,
+		fib_get_procinfo
+	});
+#endif		/* CONFIG_PROC_FS */
+
+#ifndef CONFIG_IP_MULTIPLE_TABLES
+	local_table = fib_hash_init(RT_TABLE_LOCAL);
+	main_table = fib_hash_init(RT_TABLE_MAIN);
+#else
+	fib_rules_init();
+#endif
+
+	register_netdevice_notifier(&fib_netdev_notifier);
+	register_inetaddr_notifier(&fib_inetaddr_notifier);
+}
+
diff --git a/net/ipv4/fib_hash.c b/net/ipv4/fib_hash.c
new file mode 100644
index 000000000..afa6f7fe0
--- /dev/null
+++ b/net/ipv4/fib_hash.c
@@ -0,0 +1,754 @@
+/*
+ * INET		An implementation of the TCP/IP protocol suite for the LINUX
+ *		operating system.  INET is implemented using the  BSD Socket
+ *		interface as the means of communication with the user level.
+ *
+ *		IPv4 FIB: lookup engine and maintenance routines.
+ *
+ * Version:	$Id: fib_hash.c,v 1.1 1997/11/09 19:53:13 kuznet Exp $
+ *
+ * Authors:	Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
+ *
+ *		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.
+ */
+
+#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/mm.h>
+#include <linux/string.h>
+#include <linux/socket.h>
+#include <linux/sockios.h>
+#include <linux/errno.h>
+#include <linux/in.h>
+#include <linux/inet.h>
+#include <linux/netdevice.h>
+#include <linux/if_arp.h>
+#include <linux/proc_fs.h>
+#include <linux/skbuff.h>
+#include <linux/netlink.h>
+#include <linux/init.h>
+
+#include <net/ip.h>
+#include <net/protocol.h>
+#include <net/route.h>
+#include <net/tcp.h>
+#include <net/sock.h>
+#include <net/ip_fib.h>
+
+#define FTprint(a...) 
+/*
+printk(KERN_DEBUG a)
+ */
+
+/*
+   These bizarre types are just to force strict type checking.
+   When I reversed order of bytes and changed to natural mask lengths,
+   I forgot to make fixes in several places. Now I am lazy to return
+   it back.
+ */
+
+typedef struct {
+	u32	datum;
+} fn_key_t;
+
+typedef struct {
+	u32	datum;
+} fn_hash_idx_t;
+
+struct fib_node
+{
+	struct fib_node		*fn_next;
+	struct fib_info		*fn_info;
+#define FIB_INFO(f)	((f)->fn_info)
+	fn_key_t		fn_key;
+	u8			fn_tos;
+	u8			fn_type;
+	u8			fn_scope;
+	u8			fn_state;
+};
+
+#define FN_S_ZOMBIE	1
+#define FN_S_ACCESSED	2
+
+static int fib_hash_zombies;
+
+struct fn_zone
+{
+	struct fn_zone	*fz_next;	/* Next not empty zone	*/
+	struct fib_node	**fz_hash;	/* Hash table pointer	*/
+	int		fz_nent;	/* Number of entries	*/
+
+	int		fz_divisor;	/* Hash divisor		*/
+	u32		fz_hashmask;	/* (1<<fz_divisor) - 1	*/
+#define FZ_HASHMASK(fz)	((fz)->fz_hashmask)
+
+	int		fz_order;	/* Zone order		*/
+	u32		fz_mask;
+#define FZ_MASK(fz)	((fz)->fz_mask)
+};
+
+/* NOTE. On fast computers evaluation of fz_hashmask and fz_mask
+   can be cheaper than memory lookup, so that FZ_* macros are used.
+ */
+
+struct fn_hash
+{
+	struct fn_zone	*fn_zones[33];
+	struct fn_zone	*fn_zone_list;
+};
+
+static __inline__ fn_hash_idx_t fn_hash(fn_key_t key, struct fn_zone *fz)
+{
+	u32 h = ntohl(key.datum)>>(32 - fz->fz_order);
+	h ^= (h>>20);
+	h ^= (h>>10);
+	h ^= (h>>5);
+	h &= FZ_HASHMASK(fz);
+	return *(fn_hash_idx_t*)&h;
+}
+
+#define fz_key_0(key)		((key).datum = 0)
+#define fz_prefix(key,fz)	((key).datum)
+
+static __inline__ fn_key_t fz_key(u32 dst, struct fn_zone *fz)
+{
+	fn_key_t k;
+	k.datum = dst & FZ_MASK(fz);
+	return k;
+}
+
+static __inline__ struct fib_node ** fz_chain_p(fn_key_t key, struct fn_zone *fz)
+{
+	return &fz->fz_hash[fn_hash(key, fz).datum];
+}
+
+static __inline__ struct fib_node * fz_chain(fn_key_t key, struct fn_zone *fz)
+{
+	return fz->fz_hash[fn_hash(key, fz).datum];
+}
+
+extern __inline__ int fn_key_eq(fn_key_t a, fn_key_t b)
+{
+	return a.datum == b.datum;
+}
+
+#define FZ_MAX_DIVISOR 1024
+
+#ifdef CONFIG_IP_ROUTE_LARGE_TABLES
+
+static __inline__ void fn_rebuild_zone(struct fn_zone *fz,
+					struct fib_node **old_ht,
+					int old_divisor)
+{
+	int i;
+	struct fib_node *f, **fp, *next;
+
+	for (i=0; i<old_divisor; i++) {
+		for (f=old_ht[i]; f; f=next) {
+			next = f->fn_next;
+			f->fn_next = NULL;
+			for (fp = fz_chain_p(f->fn_key, fz); *fp; fp = &(*fp)->fn_next)
+				/* NONE */;
+			*fp = f;
+		}
+	}
+}
+
+static void fn_rehash_zone(struct fn_zone *fz)
+{
+	struct fib_node **ht, **old_ht;
+	int old_divisor, new_divisor;
+	u32 new_hashmask;
+		
+	old_divisor = fz->fz_divisor;
+
+	switch (old_divisor) {
+	case 16:
+		new_divisor = 256;
+		new_hashmask = 0xFF;
+		break;
+	case 256:
+		new_divisor = 1024;
+		new_hashmask = 0x3FF;
+		break;
+	default:
+		printk(KERN_CRIT "route.c: bad divisor %d!\n", old_divisor);
+		return;
+	}
+#if RT_CACHE_DEBUG >= 2
+	printk("fn_rehash_zone: hash for zone %d grows from %d\n", fz->fz_order, old_divisor);
+#endif
+
+	ht = kmalloc(new_divisor*sizeof(struct fib_node*), GFP_KERNEL);
+
+	if (ht)	{
+		memset(ht, 0, new_divisor*sizeof(struct fib_node*));
+		start_bh_atomic();
+		old_ht = fz->fz_hash;
+		fz->fz_hash = ht;
+		fz->fz_hashmask = new_hashmask;
+		fz->fz_divisor = new_divisor;
+		fn_rebuild_zone(fz, old_ht, old_divisor);
+		end_bh_atomic();
+		kfree(old_ht);
+FTprint("REHASHED ZONE: order %d mask %08x hash %d/%08x\n", fz->fz_order, fz->fz_mask, fz->fz_divisor, fz->fz_hashmask);
+	}
+}
+#endif /* CONFIG_IP_ROUTE_LARGE_TABLES */
+
+static void fn_free_node(struct fib_node * f)
+{
+	fib_release_info(FIB_INFO(f));
+	kfree_s(f, sizeof(struct fib_node));
+}
+
+
+static struct fn_zone *
+fn_new_zone(struct fn_hash *table, int z)
+{
+	int i;
+	struct fn_zone *fz = kmalloc(sizeof(struct fn_zone), GFP_KERNEL);
+	if (!fz)
+		return NULL;
+
+	memset(fz, 0, sizeof(struct fn_zone));
+	if (z) {
+		fz->fz_divisor = 16;
+		fz->fz_hashmask = 0xF;
+	} else {
+		fz->fz_divisor = 1;
+		fz->fz_hashmask = 0;
+	}
+	fz->fz_hash = kmalloc(fz->fz_divisor*sizeof(struct fib_node*), GFP_KERNEL);
+	if (!fz->fz_hash) {
+		kfree(fz);
+		return NULL;
+	}
+	memset(fz->fz_hash, 0, fz->fz_divisor*sizeof(struct fib_node*));
+	fz->fz_order = z;
+	fz->fz_mask = inet_make_mask(z);
+
+	/* Find the first not empty zone with more specific mask */
+	for (i=z+1; i<=32; i++)
+		if (table->fn_zones[i])
+			break;
+	start_bh_atomic();
+	if (i>32) {
+		/* No more specific masks, we are the first. */
+		fz->fz_next = table->fn_zone_list;
+		table->fn_zone_list = fz;
+	} else {
+		fz->fz_next = table->fn_zones[i]->fz_next;
+		table->fn_zones[i]->fz_next = fz;
+	}
+	table->fn_zones[z] = fz;
+	end_bh_atomic();
+FTprint("NEW ZONE: order %d mask %08x hash %d/%08x\n", fz->fz_order, fz->fz_mask, fz->fz_divisor, fz->fz_hashmask);
+	return fz;
+}
+
+static int
+fn_hash_lookup(struct fib_table *tb, const struct rt_key *key, struct fib_result *res)
+{
+	int err;
+	struct fn_zone *fz;
+	struct fn_hash *t = (struct fn_hash*)tb->tb_data;
+
+	for (fz = t->fn_zone_list; fz; fz = fz->fz_next) {
+		struct fib_node *f;
+		fn_key_t k = fz_key(key->dst, fz);
+		int matched = 0;
+
+		for (f = fz_chain(k, fz); f; f = f->fn_next) {
+			if (!fn_key_eq(k, f->fn_key)
+#ifdef CONFIG_IP_ROUTE_TOS
+			    || (f->fn_tos && f->fn_tos != key->tos)
+#endif
+			    ) {
+				if (matched)
+					return 1;
+				continue;
+			}
+			matched = 1;
+			f->fn_state |= FN_S_ACCESSED;
+
+			if (f->fn_state&FN_S_ZOMBIE)
+				continue;
+			if (f->fn_scope < key->scope)
+				continue;
+
+			err = fib_semantic_match(f->fn_type, FIB_INFO(f), key, res);
+			if (err == 0) {
+				res->type = f->fn_type;
+				res->scope = f->fn_scope;
+				res->prefixlen = fz->fz_order;
+				res->prefix = &fz_prefix(f->fn_key, fz);
+				return 0;
+			}
+			if (err < 0)
+				return err;
+		}
+	}
+	return 1;
+}
+
+#define FIB_SCAN(f, fp) \
+for ( ; ((f) = *(fp)) != NULL; (fp) = &(f)->fn_next)
+
+#define FIB_SCAN_KEY(f, fp, key) \
+for ( ; ((f) = *(fp)) != NULL && fn_key_eq((f)->fn_key, (key)); (fp) = &(f)->fn_next)
+
+#define FIB_CONTINUE(f, fp) \
+{ \
+	fp = &f->fn_next; \
+	continue; \
+}
+
+#ifdef CONFIG_RTNETLINK
+static void rtmsg_fib(int, struct fib_node*, int, int,
+		      struct nlmsghdr *n,
+		      struct netlink_skb_parms *);
+#else
+#define rtmsg_fib(a, b, c, d, e, f)
+#endif
+
+
+static int
+fn_hash_insert(struct fib_table *tb, struct rtmsg *r, struct kern_rta *rta,
+		struct nlmsghdr *n, struct netlink_skb_parms *req)
+{
+	struct fn_hash *table = (struct fn_hash*)tb->tb_data;
+	struct fib_node *new_f, *f, **fp;
+	struct fn_zone *fz;
+	struct fib_info *fi;
+
+	int z = r->rtm_dst_len;
+	int type = r->rtm_type;
+#ifdef CONFIG_IP_ROUTE_TOS
+	u8 tos = r->rtm_tos;
+#endif
+	fn_key_t key;
+	unsigned state = 0;
+	int err;
+
+FTprint("tb(%d)_insert: %d %08x/%d %d %08x\n", tb->tb_id, r->rtm_type, rta->rta_dst ?
+*(u32*)rta->rta_dst : 0, z, rta->rta_oif ? *rta->rta_oif : -1,
+rta->rta_prefsrc ? *(u32*)rta->rta_prefsrc : 0);
+	if (z > 32)
+		return -EINVAL;
+	fz = table->fn_zones[z];
+	if (!fz && !(fz = fn_new_zone(table, z)))
+		return -ENOBUFS;
+
+	fz_key_0(key);
+	if (rta->rta_dst) {
+		u32 dst;
+		memcpy(&dst, rta->rta_dst, 4);
+		if (dst & ~FZ_MASK(fz))
+			return -EINVAL;
+		key = fz_key(dst, fz);
+	}
+
+	if  ((fi = fib_create_info(r, rta, n, &err)) == NULL) {
+FTprint("fib_create_info err=%d\n", err);
+		return err;
+	}
+
+#ifdef CONFIG_IP_ROUTE_LARGE_TABLES
+	if (fz->fz_nent > (fz->fz_divisor<<2) &&
+	    fz->fz_divisor < FZ_MAX_DIVISOR &&
+	    (z==32 || (1<<z) > fz->fz_divisor))
+		fn_rehash_zone(fz);
+#endif
+
+	fp = fz_chain_p(key, fz);
+
+	/*
+	 * Scan list to find the first route with the same destination
+	 */
+	FIB_SCAN(f, fp) {
+		if (fn_key_eq(f->fn_key,key))
+			break;
+	}
+
+#ifdef CONFIG_IP_ROUTE_TOS
+	/*
+	 * Find route with the same destination and tos.
+	 */
+	FIB_SCAN_KEY(f, fp, key) {
+		if (f->fn_tos <= tos)
+			break;
+	}
+#endif
+
+	if (f && fn_key_eq(f->fn_key, key)
+#ifdef CONFIG_IP_ROUTE_TOS
+	    && f->fn_tos == tos
+#endif
+	    ) {
+		state = f->fn_state;
+		if (n->nlmsg_flags&NLM_F_EXCL && !(state&FN_S_ZOMBIE))
+			return -EEXIST;
+		if (n->nlmsg_flags&NLM_F_REPLACE) {
+			struct fib_info *old_fi = FIB_INFO(f);
+			if (old_fi != fi) {
+				rtmsg_fib(RTM_DELROUTE, f, z, tb->tb_id, n, req);
+				start_bh_atomic();
+				FIB_INFO(f) = fi;
+				f->fn_type = r->rtm_type;
+				f->fn_scope = r->rtm_scope;
+				end_bh_atomic();
+				rtmsg_fib(RTM_NEWROUTE, f, z, tb->tb_id, n, req);
+			}
+			state = f->fn_state;
+			f->fn_state = 0;
+			fib_release_info(old_fi);
+			if (state&FN_S_ACCESSED)
+				rt_cache_flush(RT_FLUSH_DELAY);
+			return 0;
+		}
+		for ( ; (f = *fp) != NULL && fn_key_eq(f->fn_key, key)
+#ifdef CONFIG_IP_ROUTE_TOS
+		     && f->fn_tos == tos
+#endif
+		     ; fp = &f->fn_next) {
+			state |= f->fn_state;
+			if (f->fn_type == type && f->fn_scope == r->rtm_scope
+			    && FIB_INFO(f) == fi) {
+				fib_release_info(fi);
+				if (f->fn_state&FN_S_ZOMBIE) {
+					f->fn_state = 0;
+					rtmsg_fib(RTM_NEWROUTE, f, z, tb->tb_id, n, req);
+					if (state&FN_S_ACCESSED)
+						rt_cache_flush(RT_FLUSH_DELAY);
+					return 0;
+				}
+				return -EEXIST;
+			}
+		}
+	} else {
+		if (!(n->nlmsg_flags&NLM_F_CREATE))
+			return -ENOENT;
+	}
+
+	new_f = (struct fib_node *) kmalloc(sizeof(struct fib_node), GFP_KERNEL);
+	if (new_f == NULL) {
+		fib_release_info(fi);
+		return -ENOBUFS;
+	}
+
+	memset(new_f, 0, sizeof(struct fib_node));
+
+	new_f->fn_key = key;
+#ifdef CONFIG_IP_ROUTE_TOS
+	new_f->fn_tos = tos;
+#endif
+	new_f->fn_type = type;
+	new_f->fn_scope = r->rtm_scope;
+	FIB_INFO(new_f) = fi;
+
+	/*
+	 * Insert new entry to the list.
+	 */
+
+	start_bh_atomic();
+	new_f->fn_next = f;
+	*fp = new_f;
+	end_bh_atomic();
+	fz->fz_nent++;
+
+	rtmsg_fib(RTM_NEWROUTE, new_f, z, tb->tb_id, n, req);
+	rt_cache_flush(RT_FLUSH_DELAY);
+	return 0;
+}
+
+
+static int
+fn_hash_delete(struct fib_table *tb, struct rtmsg *r, struct kern_rta *rta,
+		struct nlmsghdr *n, struct netlink_skb_parms *req)
+{
+	struct fn_hash *table = (struct fn_hash*)tb->tb_data;
+	struct fib_node **fp, *f;
+	int z = r->rtm_dst_len;
+	struct fn_zone *fz;
+	fn_key_t key;
+#ifdef CONFIG_IP_ROUTE_TOS
+	u8 tos = r->rtm_tos;
+#endif
+
+FTprint("tb(%d)_delete: %d %08x/%d %d\n", tb->tb_id, r->rtm_type, rta->rta_dst ?
+       *(u32*)rta->rta_dst : 0, z, rta->rta_oif ? *rta->rta_oif : -1);
+	if (z > 32)
+		return -EINVAL;
+	if ((fz  = table->fn_zones[z]) == NULL)
+		return -ESRCH;
+
+	fz_key_0(key);
+	if (rta->rta_dst) {
+		u32 dst;
+		memcpy(&dst, rta->rta_dst, 4);
+		if (dst & ~FZ_MASK(fz))
+			return -EINVAL;
+		key = fz_key(dst, fz);
+	}
+
+	fp = fz_chain_p(key, fz);
+
+	FIB_SCAN(f, fp) {
+		if (fn_key_eq(f->fn_key, key))
+			break;
+	}
+#ifdef CONFIG_IP_ROUTE_TOS
+	FIB_SCAN_KEY(f, fp, key) {
+		if (f->fn_tos == tos)
+			break;
+	}
+#endif
+
+	while ((f = *fp) != NULL && fn_key_eq(f->fn_key, key)
+#ifdef CONFIG_IP_ROUTE_TOS
+	       && f->fn_tos == tos
+#endif
+	       ) {
+		struct fib_info * fi = FIB_INFO(f);
+
+		if ((f->fn_state&FN_S_ZOMBIE) ||
+		    (r->rtm_type && f->fn_type != r->rtm_type) ||
+		    (r->rtm_scope && f->fn_scope != r->rtm_scope) ||
+		    (r->rtm_protocol && fi->fib_protocol != r->rtm_protocol) ||
+		    fib_nh_match(r, n, rta, fi))
+			FIB_CONTINUE(f, fp);
+		break;
+	}
+	if (!f)
+		return -ESRCH;
+#if 0
+	*fp = f->fn_next;
+	rtmsg_fib(RTM_DELROUTE, f, z, tb->tb_id, n, req);
+	fn_free_node(f);
+	fz->fz_nent--;
+	rt_cache_flush(0);
+#else
+	f->fn_state |= FN_S_ZOMBIE;
+	rtmsg_fib(RTM_DELROUTE, f, z, tb->tb_id, n, req);
+	if (f->fn_state&FN_S_ACCESSED) {
+		f->fn_state &= ~FN_S_ACCESSED;
+		rt_cache_flush(RT_FLUSH_DELAY);
+	}
+	if (++fib_hash_zombies > 128)
+		fib_flush();
+#endif
+	return 0;
+}
+
+extern __inline__ int
+fn_flush_list(struct fib_node ** fp, int z, struct fn_hash *table)
+{
+	int found = 0;
+	struct fib_node *f;
+
+	while ((f = *fp) != NULL) {
+		struct fib_info *fi = FIB_INFO(f);
+
+		if (fi && ((f->fn_state&FN_S_ZOMBIE) || (fi->fib_flags&RTNH_F_DEAD))) {
+			*fp = f->fn_next;
+			fn_free_node(f);
+			found++;
+			continue;
+		}
+		fp = &f->fn_next;
+	}
+	return found;
+}
+
+static int fn_hash_flush(struct fib_table *tb)
+{
+	struct fn_hash *table = (struct fn_hash*)tb->tb_data;
+	struct fn_zone *fz;
+	int found = 0;
+
+	fib_hash_zombies = 0;
+	for (fz = table->fn_zone_list; fz; fz = fz->fz_next) {
+		int i;
+		int tmp = 0;
+		for (i=fz->fz_divisor-1; i>=0; i--)
+			tmp += fn_flush_list(&fz->fz_hash[i], fz->fz_order, table);
+		fz->fz_nent -= tmp;
+		found += tmp;
+	}
+	return found;
+}
+
+
+#ifdef CONFIG_PROC_FS
+
+static int fn_hash_get_info(struct fib_table *tb, char *buffer, int first, int count)
+{
+	struct fn_hash *table = (struct fn_hash*)tb->tb_data;
+	struct fn_zone *fz;
+	int pos = 0;
+	int n = 0;
+
+	for (fz=table->fn_zone_list; fz; fz = fz->fz_next) {
+		int i;
+		struct fib_node *f;
+		int maxslot = fz->fz_divisor;
+		struct fib_node **fp = fz->fz_hash;
+
+		if (fz->fz_nent == 0)
+			continue;
+
+		if (pos + fz->fz_nent <= first) {
+			pos += fz->fz_nent;
+			continue;
+		}
+
+		for (i=0; i < maxslot; i++, fp++) {
+			for (f = *fp; f; f = f->fn_next) {
+				if (++pos <= first)
+					continue;
+				fib_node_get_info(f->fn_type,
+						  f->fn_state&FN_S_ZOMBIE,
+						  FIB_INFO(f),
+						  fz_prefix(f->fn_key, fz),
+						  FZ_MASK(fz), buffer);
+				buffer += 128;
+				if (++n >= count)
+					return n;
+			}
+		}
+	}
+  	return n;
+}
+#endif
+
+
+#ifdef CONFIG_RTNETLINK
+
+extern __inline__ int
+fn_hash_dump_bucket(struct sk_buff *skb, struct netlink_callback *cb,
+		     struct fib_table *tb,
+		     struct fn_zone *fz,
+		     struct fib_node *f)
+{
+	int i, s_i;
+
+	s_i = cb->args[3];
+	for (i=0; f; i++, f=f->fn_next) {
+		if (i < s_i) continue;
+		if (f->fn_state&FN_S_ZOMBIE) continue;
+		if (fib_dump_info(skb, NETLINK_CB(cb->skb).pid, cb->nlh->nlmsg_seq,
+				  RTM_NEWROUTE,
+				  tb->tb_id, (f->fn_state&FN_S_ZOMBIE) ? 0 : f->fn_type, f->fn_scope,
+				  &f->fn_key, fz->fz_order, f->fn_tos,
+				  f->fn_info) < 0) {
+			cb->args[3] = i;
+			return -1;
+		}
+	}
+	cb->args[3] = i;
+	return skb->len;
+}
+
+extern __inline__ int
+fn_hash_dump_zone(struct sk_buff *skb, struct netlink_callback *cb,
+		   struct fib_table *tb,
+		   struct fn_zone *fz)
+{
+	int h, s_h;
+
+	s_h = cb->args[2];
+	for (h=0; h < fz->fz_divisor; h++) {
+		if (h < s_h) continue;
+		if (h > s_h)
+			memset(&cb->args[3], 0, sizeof(cb->args) - 3*sizeof(int));
+		if (fz->fz_hash == NULL || fz->fz_hash[h] == NULL)
+			continue;
+		if (fn_hash_dump_bucket(skb, cb, tb, fz, fz->fz_hash[h]) < 0) {
+			cb->args[2] = h;
+			return -1;
+		}
+	}
+	cb->args[2] = h;
+	return skb->len;
+}
+
+static int fn_hash_dump(struct fib_table *tb, struct sk_buff *skb, struct netlink_callback *cb)
+{
+	int m, s_m;
+	struct fn_zone *fz;
+	struct fn_hash *table = (struct fn_hash*)tb->tb_data;
+
+	s_m = cb->args[1];
+	for (fz = table->fn_zone_list, m=0; fz; fz = fz->fz_next, m++) {
+		if (m < s_m) continue;
+		if (m > s_m)
+			memset(&cb->args[2], 0, sizeof(cb->args) - 2*sizeof(int));
+		if (fn_hash_dump_zone(skb, cb, tb, fz) < 0) {
+			cb->args[1] = m;
+			return -1;
+		}
+	}
+	cb->args[1] = m;
+	return skb->len;
+}
+
+static void rtmsg_fib(int event, struct fib_node* f, int z, int tb_id,
+		      struct nlmsghdr *n, struct netlink_skb_parms *req)
+{
+	struct sk_buff *skb;
+	pid_t pid = req ? req->pid : 0;
+	int size = NLMSG_SPACE(sizeof(struct rtmsg)+256);
+
+	skb = alloc_skb(size, GFP_KERNEL);
+	if (!skb)
+		return;
+
+	if (fib_dump_info(skb, pid, n->nlmsg_seq, event, tb_id,
+			  f->fn_type, f->fn_scope, &f->fn_key, z, f->fn_tos,
+			  FIB_INFO(f)) < 0) {
+		kfree_skb(skb, 0);
+		return;
+	}
+	NETLINK_CB(skb).dst_groups = RTMGRP_IPV4_ROUTE;
+	if (n->nlmsg_flags&NLM_F_ECHO)
+		atomic_inc(&skb->users);
+	netlink_broadcast(rtnl, skb, pid, RTMGRP_IPV4_ROUTE, GFP_KERNEL);
+	if (n->nlmsg_flags&NLM_F_ECHO)
+		netlink_unicast(rtnl, skb, pid, MSG_DONTWAIT);
+}
+
+#endif /* CONFIG_RTNETLINK */
+
+#ifdef CONFIG_IP_MULTIPLE_TABLES
+struct fib_table * fib_hash_init(int id)
+#else
+__initfunc(struct fib_table * fib_hash_init(int id))
+#endif
+{
+	struct fib_table *tb;
+	tb = kmalloc(sizeof(struct fib_table) + sizeof(struct fn_hash), GFP_KERNEL);
+	if (tb == NULL)
+		return NULL;
+	tb->tb_id = id;
+	tb->tb_lookup = fn_hash_lookup;
+	tb->tb_insert = fn_hash_insert;
+	tb->tb_delete = fn_hash_delete;
+	tb->tb_flush = fn_hash_flush;
+#ifdef CONFIG_RTNETLINK
+	tb->tb_dump = fn_hash_dump;
+#endif
+#ifdef CONFIG_PROC_FS
+	tb->tb_get_info = fn_hash_get_info;
+#endif
+	memset(tb->tb_data, 0, sizeof(struct fn_hash));
+	return tb;
+}
diff --git a/net/ipv4/fib_rules.c b/net/ipv4/fib_rules.c
new file mode 100644
index 000000000..c593d758f
--- /dev/null
+++ b/net/ipv4/fib_rules.c
@@ -0,0 +1,363 @@
+/*
+ * INET		An implementation of the TCP/IP protocol suite for the LINUX
+ *		operating system.  INET is implemented using the  BSD Socket
+ *		interface as the means of communication with the user level.
+ *
+ *		IPv4 Forwarding Information Base: policy rules.
+ *
+ * Version:	$Id: fib_rules.c,v 1.2 1997/10/10 22:40:49 davem Exp $
+ *
+ * Authors:	Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
+ *
+ *		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.
+ */
+
+#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/mm.h>
+#include <linux/string.h>
+#include <linux/socket.h>
+#include <linux/sockios.h>
+#include <linux/errno.h>
+#include <linux/in.h>
+#include <linux/inet.h>
+#include <linux/netdevice.h>
+#include <linux/if_arp.h>
+#include <linux/proc_fs.h>
+#include <linux/skbuff.h>
+#include <linux/netlink.h>
+#include <linux/init.h>
+
+#include <net/ip.h>
+#include <net/protocol.h>
+#include <net/route.h>
+#include <net/tcp.h>
+#include <net/sock.h>
+#include <net/ip_fib.h>
+
+#define FRprintk(a...)
+
+struct fib_rule
+{
+	struct fib_rule *r_next;
+	unsigned	r_preference;
+	unsigned char	r_table;
+	unsigned char	r_action;
+	unsigned char	r_dst_len;
+	unsigned char	r_src_len;
+	u32		r_src;
+	u32		r_srcmask;
+	u32		r_dst;
+	u32		r_dstmask;
+	u32		r_srcmap;
+	u8		r_flags;
+	u8		r_tos;
+	int		r_ifindex;
+	char		r_ifname[IFNAMSIZ];
+};
+
+static struct fib_rule default_rule = { NULL, 0x7FFF, RT_TABLE_DEFAULT, RTN_UNICAST, };
+static struct fib_rule main_rule = { &default_rule, 0x7FFE, RT_TABLE_MAIN, RTN_UNICAST, };
+static struct fib_rule local_rule = { &main_rule, 0, RT_TABLE_LOCAL, RTN_UNICAST, };
+
+static struct fib_rule *fib_rules = &local_rule;
+
+int inet_rtm_delrule(struct sk_buff *skb, struct nlmsghdr* nlh, void *arg)
+{
+	struct kern_rta *rta = arg;
+	struct rtmsg *rtm = NLMSG_DATA(nlh);
+	struct fib_rule *r, **rp;
+
+	for (rp=&fib_rules; (r=*rp) != NULL; rp=&r->r_next) {
+		if ((!rta->rta_src || memcmp(rta->rta_src, &r->r_src, 4) == 0) &&
+		    rtm->rtm_src_len == r->r_src_len &&
+		    rtm->rtm_dst_len == r->r_dst_len &&
+		    (!rta->rta_dst || memcmp(rta->rta_dst, &r->r_dst, 4) == 0) &&
+		    rtm->rtm_tos == r->r_tos &&
+		    rtm->rtm_type == r->r_action &&
+		    (!rta->rta_priority || *rta->rta_priority == r->r_preference) &&
+		    (!rta->rta_ifname || strcmp(rta->rta_ifname, r->r_ifname) == 0) &&
+		    (!rtm->rtm_table || (r && rtm->rtm_table == r->r_table))) {
+			*rp = r->r_next;
+			if (r != &default_rule && r != &main_rule && r != &local_rule)
+				kfree(r);
+			return 0;
+		}
+	}
+	return -ESRCH;
+}
+
+/* Allocate new unique table id */
+
+static struct fib_table *fib_empty_table(void)
+{
+	int id;
+
+	for (id = 1; id <= RT_TABLE_MAX; id++)
+		if (fib_tables[id] == NULL)
+			return __fib_new_table(id);
+	return NULL;
+}
+
+
+int inet_rtm_newrule(struct sk_buff *skb, struct nlmsghdr* nlh, void *arg)
+{
+	struct kern_rta *rta = arg;
+	struct rtmsg *rtm = NLMSG_DATA(nlh);
+	struct fib_rule *r, *new_r, **rp;
+	unsigned char table_id;
+
+	if (rtm->rtm_src_len > 32 || rtm->rtm_dst_len > 32 ||
+	    (rtm->rtm_tos & ~IPTOS_TOS_MASK))
+		return -EINVAL;
+
+	table_id = rtm->rtm_table;
+	if (table_id == RT_TABLE_UNSPEC) {
+		struct fib_table *table;
+		if (rtm->rtm_type == RTN_UNICAST || rtm->rtm_type == RTN_NAT) {
+			if ((table = fib_empty_table()) == NULL)
+				return -ENOBUFS;
+			table_id = table->tb_id;
+		}
+	}
+
+	new_r = kmalloc(sizeof(*new_r), GFP_KERNEL);
+	if (!new_r)
+		return -ENOMEM;
+	memset(new_r, 0, sizeof(*new_r));
+	if (rta->rta_src)
+		memcpy(&new_r->r_src, rta->rta_src, 4);
+	if (rta->rta_dst)
+		memcpy(&new_r->r_dst, rta->rta_dst, 4);
+	if (rta->rta_gw)
+		memcpy(&new_r->r_srcmap, rta->rta_gw, 4);
+	new_r->r_src_len = rtm->rtm_src_len;
+	new_r->r_dst_len = rtm->rtm_dst_len;
+	new_r->r_srcmask = inet_make_mask(rtm->rtm_src_len);
+	new_r->r_dstmask = inet_make_mask(rtm->rtm_dst_len);
+	new_r->r_tos = rtm->rtm_tos;
+	new_r->r_action = rtm->rtm_type;
+	new_r->r_flags = rtm->rtm_flags;
+	if (rta->rta_priority)
+		new_r->r_preference = *rta->rta_priority;
+	new_r->r_table = table_id;
+	if (rta->rta_ifname) {
+		struct device *dev;
+		memcpy(new_r->r_ifname, rta->rta_ifname, IFNAMSIZ);
+		new_r->r_ifindex = -1;
+		dev = dev_get(rta->rta_ifname);
+		if (dev)
+			new_r->r_ifindex = dev->ifindex;
+	}
+
+	rp = &fib_rules;
+	if (!new_r->r_preference) {
+		r = fib_rules;
+		if (r && (r = r->r_next) != NULL) {
+			rp = &fib_rules->r_next;
+			if (r->r_preference)
+				new_r->r_preference = r->r_preference - 1;
+		}
+	}
+
+	while ( (r = *rp) != NULL ) {
+		if (r->r_preference > new_r->r_preference)
+			break;
+		rp = &r->r_next;
+	}
+
+	new_r->r_next = r;
+	*rp = new_r;
+	return 0;
+}
+
+u32 fib_rules_map_destination(u32 daddr, struct fib_result *res)
+{
+	u32 mask = inet_make_mask(res->prefixlen);
+	return (daddr&~mask)|res->fi->fib_nh->nh_gw;
+}
+
+u32 fib_rules_policy(u32 saddr, struct fib_result *res, unsigned *flags)
+{
+	struct fib_rule *r = res->r;
+
+	if (r->r_action == RTN_NAT) {
+		int addrtype = inet_addr_type(r->r_srcmap);
+
+		if (addrtype == RTN_NAT) {
+			/* Packet is from  translated source; remember it */
+			saddr = (saddr&~r->r_srcmask)|r->r_srcmap;
+			*flags |= RTCF_SNAT;
+		} else if (addrtype == RTN_LOCAL || r->r_srcmap == 0) {
+			/* Packet is from masqueraded source; remember it */
+			saddr = r->r_srcmap;
+			*flags |= RTCF_MASQ;
+		}
+	}
+	return saddr;
+}
+
+static void fib_rules_detach(struct device *dev)
+{
+	struct fib_rule *r;
+
+	for (r=fib_rules; r; r=r->r_next) {
+		if (r->r_ifindex == dev->ifindex)
+			r->r_ifindex = -1;
+	}
+}
+
+static void fib_rules_attach(struct device *dev)
+{
+	struct fib_rule *r;
+
+	for (r=fib_rules; r; r=r->r_next) {
+		if (r->r_ifindex == -1 && strcmp(dev->name, r->r_ifname) == 0)
+			r->r_ifindex = dev->ifindex;
+	}
+}
+
+int fib_lookup(const struct rt_key *key, struct fib_result *res)
+{
+	int err;
+	struct fib_rule *r, *policy;
+	struct fib_table *tb;
+
+	u32 daddr = key->dst;
+	u32 saddr = key->src;
+
+FRprintk("Lookup: %08x <- %08x ", key->dst, key->src);
+	for (r = fib_rules; r; r=r->r_next) {
+		if (((saddr^r->r_src) & r->r_srcmask) ||
+		    ((daddr^r->r_dst) & r->r_dstmask) ||
+#ifdef CONFIG_IP_TOS_ROUTING
+		    (r->r_tos && r->r_tos != key->tos) ||
+#endif
+		    (r->r_ifindex && r->r_ifindex != key->iif))
+			continue;
+
+FRprintk("tb %d r %d ", r->r_table, r->r_action);
+		switch (r->r_action) {
+		case RTN_UNICAST:
+			policy = NULL;
+			break;
+		case RTN_NAT:
+			policy = r;
+			break;
+		case RTN_UNREACHABLE:
+			return -ENETUNREACH;
+		default:
+		case RTN_BLACKHOLE:
+			return -EINVAL;
+		case RTN_PROHIBIT:
+			return -EACCES;
+		}
+
+		if ((tb = fib_get_table(r->r_table)) == NULL)
+			continue;
+		err = tb->tb_lookup(tb, key, res);
+		if (err == 0) {
+FRprintk("ok\n");
+			res->r = policy;
+			return 0;
+		}
+		if (err < 0)
+			return err;
+FRprintk("RCONT ");
+	}
+FRprintk("FAILURE\n");
+	return -ENETUNREACH;
+}
+
+static int fib_rules_event(struct notifier_block *this, unsigned long event, void *ptr)
+{
+	struct device *dev = ptr;
+
+	if (event == NETDEV_UNREGISTER)
+		fib_rules_detach(dev);
+	else if (event == NETDEV_REGISTER)
+		fib_rules_attach(dev);
+	return NOTIFY_DONE;
+}
+
+
+struct notifier_block fib_rules_notifier = {
+	fib_rules_event,
+	NULL,
+	0
+};
+
+#ifdef CONFIG_RTNETLINK
+
+extern __inline__ int inet_fill_rule(struct sk_buff *skb,
+				     struct fib_rule *r,
+				     struct netlink_callback *cb)
+{
+	struct rtmsg *rtm;
+	struct nlmsghdr  *nlh;
+	unsigned char	 *b = skb->tail;
+
+	nlh = NLMSG_PUT(skb, NETLINK_CREDS(cb->skb)->pid, cb->nlh->nlmsg_seq, RTM_NEWRULE, sizeof(*rtm));
+	rtm = NLMSG_DATA(nlh);
+	rtm->rtm_family = AF_INET;
+	rtm->rtm_dst_len = r->r_dst_len;
+	rtm->rtm_src_len = r->r_src_len;
+	rtm->rtm_tos = r->r_tos;
+	rtm->rtm_table = r->r_table;
+	rtm->rtm_protocol = 0;
+	rtm->rtm_scope = 0;
+	rtm->rtm_nhs = 0;
+	rtm->rtm_type = r->r_action;
+	rtm->rtm_optlen = 0;
+	rtm->rtm_flags = r->r_flags;
+
+	if (r->r_dst_len)
+		RTA_PUT(skb, RTA_DST, 4, &r->r_dst);
+	if (r->r_src_len)
+		RTA_PUT(skb, RTA_SRC, 4, &r->r_src);
+	if (r->r_ifname[0])
+		RTA_PUT(skb, RTA_IFNAME, IFNAMSIZ, &r->r_ifname);
+	if (r->r_preference)
+		RTA_PUT(skb, RTA_PRIORITY, 4, &r->r_preference);
+	if (r->r_srcmap)
+		RTA_PUT(skb, RTA_GATEWAY, 4, &r->r_srcmap);
+	nlh->nlmsg_len = skb->tail - b;
+	return skb->len;
+
+nlmsg_failure:
+rtattr_failure:
+	skb_put(skb, b - skb->tail);
+	return -1;
+}
+
+int inet_dump_rules(struct sk_buff *skb, struct netlink_callback *cb)
+{
+	int idx;
+	int s_idx = cb->args[0];
+	struct fib_rule *r;
+
+	for (r=fib_rules, idx=0; r; r = r->r_next, idx++) {
+		if (idx < s_idx)
+			continue;
+		if (inet_fill_rule(skb, r, cb) < 0)
+			break;
+	}
+	cb->args[0] = idx;
+
+	return skb->len;
+}
+
+#endif /* CONFIG_RTNETLINK */
+
+__initfunc(void fib_rules_init(void))
+{
+	register_netdevice_notifier(&fib_rules_notifier);
+}
diff --git a/net/ipv4/fib_semantics.c b/net/ipv4/fib_semantics.c
new file mode 100644
index 000000000..8f3e70cad
--- /dev/null
+++ b/net/ipv4/fib_semantics.c
@@ -0,0 +1,908 @@
+/*
+ * INET		An implementation of the TCP/IP protocol suite for the LINUX
+ *		operating system.  INET is implemented using the  BSD Socket
+ *		interface as the means of communication with the user level.
+ *
+ *		IPv4 Forwarding Information Base: semantics.
+ *
+ * Version:	$Id: fib_semantics.c,v 1.5 1997/10/10 22:40:50 davem Exp $
+ *
+ * Authors:	Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
+ *
+ *		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.
+ */
+
+#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/mm.h>
+#include <linux/string.h>
+#include <linux/socket.h>
+#include <linux/sockios.h>
+#include <linux/errno.h>
+#include <linux/in.h>
+#include <linux/inet.h>
+#include <linux/netdevice.h>
+#include <linux/if_arp.h>
+#include <linux/proc_fs.h>
+#include <linux/skbuff.h>
+#include <linux/netlink.h>
+#include <linux/init.h>
+
+#include <net/ip.h>
+#include <net/protocol.h>
+#include <net/route.h>
+#include <net/tcp.h>
+#include <net/sock.h>
+#include <net/ip_fib.h>
+
+#define FSprintk(a...)
+
+static struct fib_info 	*fib_info_list;
+
+#define for_fib_info() { struct fib_info *fi; \
+	for (fi = fib_info_list; fi; fi = fi->fib_next)
+
+#define endfor_fib_info() }
+
+#ifdef CONFIG_IP_ROUTE_MULTIPATH
+
+#define for_nexthops(fi) { int nhsel; const struct fib_nh * nh; \
+for (nhsel=0, nh = (fi)->fib_nh; nhsel < (fi)->fib_nhs; nh++, nhsel++)
+
+#define change_nexthops(fi) { int nhsel; struct fib_nh * nh; \
+for (nhsel=0, nh = (struct fib_nh*)((fi)->fib_nh); nhsel < (fi)->fib_nhs; nh++, nhsel++)
+
+#else /* CONFIG_IP_ROUTE_MULTIPATH */
+
+/* Hope, that gcc will optimize it to get rid of dummy loop */
+
+#define for_nexthops(fi) { int nhsel=0; const struct fib_nh * nh = (fi)->fib_nh; \
+for (nhsel=0; nhsel < 1; nhsel++)
+
+#define change_nexthops(fi) { int nhsel=0; struct fib_nh * nh = (struct fib_nh*)((fi)->fib_nh); \
+for (nhsel=0; nhsel < 1; nhsel++)
+
+#endif /* CONFIG_IP_ROUTE_MULTIPATH */
+
+#define endfor_nexthops(fi) }
+
+
+static struct 
+{
+	int	error;
+	u8	scope;
+} fib_props[RTA_MAX+1] = {
+        { 0, RT_SCOPE_NOWHERE},		/* RTN_UNSPEC */
+	{ 0, RT_SCOPE_UNIVERSE},	/* RTN_UNICAST */
+	{ 0, RT_SCOPE_HOST},		/* RTN_LOCAL */
+	{ 0, RT_SCOPE_LINK},		/* RTN_BROADCAST */
+	{ 0, RT_SCOPE_LINK},		/* RTN_ANYCAST */
+	{ 0, RT_SCOPE_UNIVERSE},	/* RTN_MULTICAST */
+	{ -EINVAL, RT_SCOPE_UNIVERSE},	/* RTN_BLACKHOLE */
+	{ -EHOSTUNREACH, RT_SCOPE_UNIVERSE},/* RTN_UNREACHABLE */
+	{ -EACCES, RT_SCOPE_UNIVERSE},	/* RTN_PROHIBIT */
+	{ 1, RT_SCOPE_UNIVERSE},	/* RTN_THROW */
+#ifdef CONFIG_IP_ROUTE_NAT
+	{ 0, RT_SCOPE_HOST},		/* RTN_NAT */
+#else
+	{ -EINVAL, RT_SCOPE_NOWHERE},	/* RTN_NAT */
+#endif
+	{ -EINVAL, RT_SCOPE_NOWHERE}	/* RTN_XRESOLVE */
+};
+
+/* Release a nexthop info record */
+
+void fib_release_info(struct fib_info *fi)
+{
+	if (fi && !--fi->fib_refcnt) {
+		if (fi->fib_next)
+			fi->fib_next->fib_prev = fi->fib_prev;
+		if (fi->fib_prev)
+			fi->fib_prev->fib_next = fi->fib_next;
+		if (fi == fib_info_list)
+			fib_info_list = fi->fib_next;
+		kfree(fi);
+	}
+}
+
+extern __inline__ int nh_comp(const struct fib_info *fi, const struct fib_info *ofi)
+{
+	const struct fib_nh *onh = ofi->fib_nh;
+
+	for_nexthops(fi) {
+		if (nh->nh_oif != onh->nh_oif ||
+		    nh->nh_gw  != onh->nh_gw ||
+#ifdef CONFIG_IP_ROUTE_MULTIPATH
+		    nh->nh_weight != onh->nh_weight ||
+#endif
+		    ((nh->nh_flags^onh->nh_flags)&~RTNH_F_DEAD))
+			return -1;
+		onh++;
+	} endfor_nexthops(fi);
+	return 0;
+}
+
+extern __inline__ struct fib_info * fib_find_info(const struct fib_info *nfi)
+{
+	for_fib_info() {
+		if (fi->fib_nhs != nfi->fib_nhs)
+			continue;
+		if (nfi->fib_protocol == fi->fib_protocol &&
+		    nfi->fib_prefsrc == fi->fib_prefsrc &&
+		    nfi->fib_mtu == fi->fib_mtu &&
+		    nfi->fib_rtt == fi->fib_rtt &&
+		    nfi->fib_window == fi->fib_window &&
+		    ((nfi->fib_flags^fi->fib_flags)&~RTNH_F_DEAD) == 0 &&
+		    (nfi->fib_nhs == 0 || nh_comp(fi, nfi) == 0))
+			return fi;
+	} endfor_fib_info();
+	return NULL;
+}
+
+/* Check, that the gateway is already configured.
+   Used only by redirect accept routine.
+ */
+
+int ip_fib_check_default(u32 gw, struct device *dev)
+{
+	for_fib_info() {
+		if (fi->fib_flags & RTNH_F_DEAD)
+			continue;
+		for_nexthops(fi) {
+			if (nh->nh_dev == dev && nh->nh_gw == gw &&
+			    !(nh->nh_flags&RTNH_F_DEAD))
+				return 0;
+		} endfor_nexthops(fi);
+	} endfor_fib_info();
+	return -1;
+}
+
+#ifdef CONFIG_IP_ROUTE_MULTIPATH
+
+static u32 fib_get_attr32(struct rtattr *attr, int attrlen, int type)
+{
+	while (RTA_OK(attr,attrlen)) {
+		if (attr->rta_type == type)
+			return *(u32*)RTA_DATA(attr);
+		attr = RTA_NEXT(attr, attrlen);
+	}
+	return 0;
+}
+
+static int
+fib_get_nhs(struct fib_info *fi, const struct nlmsghdr *nlh, const struct rtmsg *r)
+{
+	struct rtnexthop *nhp = RTM_RTNH(r);
+	int nhlen = RTM_NHLEN(nlh, r);
+
+printk("get nhs %d/%d\n", r->rtm_nhs, nhlen);	
+	change_nexthops(fi) {
+		int attrlen = nhlen - sizeof(struct rtnexthop);
+		if (attrlen < 0 || (nhlen -= nhp->rtnh_len) < 0)
+			return -EINVAL;
+		nh->nh_flags = (r->rtm_flags&~0xFF) | nhp->rtnh_flags;
+		nh->nh_oif = nhp->rtnh_ifindex;
+		nh->nh_weight = nhp->rtnh_hops + 1;
+		if (attrlen)
+			nh->nh_gw = fib_get_attr32(RTNH_DATA(nhp), attrlen, RTA_GATEWAY);
+printk("Got nh: via %08x dev %d w %d fl %02x\n", nh->nh_gw, nh->nh_oif,
+       nh->nh_weight, nh->nh_flags);
+		nhp = RTNH_NEXT(nhp);
+	} endfor_nexthops(fi);
+	return 0;
+}
+
+#endif
+
+int fib_nh_match(struct rtmsg *r, struct nlmsghdr *nlh, struct kern_rta *rta,
+		 struct fib_info *fi)
+{
+#ifdef CONFIG_IP_ROUTE_MULTIPATH
+	struct rtnexthop *nhp;
+	int nhlen;
+#endif
+
+	if (rta->rta_oif || rta->rta_gw) {
+		if ((!rta->rta_oif || *rta->rta_oif == fi->fib_nh->nh_oif) &&
+		    (!rta->rta_gw  || memcmp(rta->rta_gw, &fi->fib_nh->nh_gw, 4) == 0))
+			return 0;
+		return 1;
+	}
+
+#ifdef CONFIG_IP_ROUTE_MULTIPATH
+	if (r->rtm_nhs == 0)
+		return 0;
+
+	nhp = RTM_RTNH(r);
+	nhlen = RTM_NHLEN(nlh, r);
+	
+	for_nexthops(fi) {
+		int attrlen = nhlen - sizeof(struct rtnexthop);
+		u32 gw;
+
+		if (attrlen < 0 || (nhlen -= nhp->rtnh_len) < 0)
+			return -EINVAL;
+		if (nhp->rtnh_ifindex && nhp->rtnh_ifindex != nh->nh_oif)
+			return 1;
+		if (attrlen) {
+			gw = fib_get_attr32(RTNH_DATA(nhp), attrlen, RTA_GATEWAY);
+			if (gw && gw != nh->nh_gw)
+				return 1;
+		}
+		nhp = RTNH_NEXT(nhp);
+	} endfor_nexthops(fi);
+#endif
+	return 0;
+}
+
+
+/*
+   Picture
+   -------
+
+   Semantics of nexthop is very messy by historical reasons.
+   We have to take into account, that:
+   a) gateway can be actually local interface address,
+      so that gatewayed route is direct.
+   b) gateway must be on-link address, possibly
+      described not by an ifaddr, but also by a direct route.
+   c) If both gateway and interface are specified, they should not
+      contradict.
+   d) If we use tunnel routes, gateway could be not on-link.
+
+   Attempt to reconcile all of these (alas, self-contradictory) conditions
+   results in pretty ugly and hairy code with obscure logic.
+
+   I choosed to generalized it instead, so that the size
+   of code does not increase practically, but it becomes
+   much more general.
+   Every prefix is assigned a "scope" value: "host" is local address,
+   "link" is direct route,
+   [ ... "site" ... "interior" ... ]
+   and "universe" is true gateway route with global meaning.
+
+   Every prefix refers to a set of "nexthop"s (gw, oif),
+   where gw must have narrower scope. This recursion stops
+   when gw has LOCAL scope or if "nexthop" is declared ONLINK,
+   which means that gw is forced to be on link.
+
+   Code is still hairy, but now it is apparently logically
+   consistent and very flexible. F.e. as by-product it allows
+   to co-exists in peace independent exterior and interior
+   routing processes.
+
+   Normally it looks as following.
+
+   {universe prefix}  -> (gw, oif) [scope link]
+                          |
+			  |-> {link prefix} -> (gw, oif) [scope local]
+			                        |
+						|-> {local prefix} (terminal node)
+ */
+
+static int fib_check_nh(const struct rtmsg *r, struct fib_info *fi, struct fib_nh *nh)
+{
+	int err;
+
+	if (nh->nh_gw) {
+		struct rt_key key;
+		struct fib_result res;
+
+#ifdef CONFIG_IP_ROUTE_PERVASIVE
+		if (nh->nh_flags&RTNH_F_PERVASIVE)
+			return 0;
+#endif
+		if (nh->nh_flags&RTNH_F_ONLINK) {
+			struct device *dev;
+
+			if (r->rtm_scope >= RT_SCOPE_LINK)
+				return -EINVAL;
+			if (inet_addr_type(nh->nh_gw) != RTN_UNICAST)
+				return -EINVAL;
+			if ((dev = dev_get_by_index(nh->nh_oif)) == NULL)
+				return -ENODEV;
+			if (!(dev->flags&IFF_UP))
+				return -ENETDOWN;
+			nh->nh_dev = dev;
+			nh->nh_scope = RT_SCOPE_LINK;
+			return 0;
+		}
+		memset(&key, 0, sizeof(key));
+		key.dst = nh->nh_gw;
+		key.oif = nh->nh_oif;
+		key.scope = r->rtm_scope + 1;
+
+		/* It is not necessary, but requires a bit of thinking */
+		if (key.scope < RT_SCOPE_LINK)
+			key.scope = RT_SCOPE_LINK;
+
+		if ((err = fib_lookup(&key, &res)) != 0)
+			return err;
+		nh->nh_scope = res.scope;
+		nh->nh_oif = FIB_RES_OIF(res);
+		nh->nh_dev = FIB_RES_DEV(res);
+	} else {
+		struct in_device *in_dev;
+
+		if (nh->nh_flags&(RTNH_F_PERVASIVE|RTNH_F_ONLINK))
+			return -EINVAL;
+
+		in_dev = inetdev_by_index(nh->nh_oif);
+		if (in_dev == NULL)
+			return -ENODEV;
+		if (!(in_dev->dev->flags&IFF_UP))
+			return -ENETDOWN;
+		nh->nh_dev = in_dev->dev;
+		nh->nh_scope = RT_SCOPE_HOST;
+	}
+	return 0;
+}
+
+struct fib_info *
+fib_create_info(const struct rtmsg *r, struct kern_rta *rta,
+		const struct nlmsghdr *nlh, int *errp)
+{
+	int err;
+	struct fib_info *fi = NULL;
+	struct fib_info *ofi;
+#ifdef CONFIG_IP_ROUTE_MULTIPATH
+	int nhs = r->rtm_nhs ? : 1;
+#else
+	const int nhs = 1;
+#endif
+
+	/* Fast check to catch the most weird cases */
+	if (fib_props[r->rtm_type].scope > r->rtm_scope) {
+		printk("Einval 1\n");
+		goto err_inval;
+	}
+
+	fi = kmalloc(sizeof(*fi)+nhs*sizeof(struct fib_nh), GFP_KERNEL);
+	err = -ENOBUFS;
+	if (fi == NULL)
+		goto failure;
+	memset(fi, 0, sizeof(*fi)+nhs*sizeof(struct fib_nh));
+
+	fi->fib_protocol = r->rtm_protocol;
+	fi->fib_nhs = nhs;
+	fi->fib_flags = r->rtm_flags;
+	if (rta->rta_mtu)
+		fi->fib_mtu = *rta->rta_mtu;
+	if (rta->rta_rtt)
+		fi->fib_rtt = *rta->rta_rtt;
+	if (rta->rta_window)
+		fi->fib_window = *rta->rta_window;
+	if (rta->rta_prefsrc)
+		memcpy(&fi->fib_prefsrc, rta->rta_prefsrc, 4);
+
+	if (r->rtm_nhs) {
+#ifdef CONFIG_IP_ROUTE_MULTIPATH
+		if ((err = fib_get_nhs(fi, nlh, r)) != 0)
+			goto failure;
+		if (rta->rta_oif && fi->fib_nh->nh_oif != *rta->rta_oif)
+			goto err_inval;
+		if (rta->rta_gw && memcmp(&fi->fib_nh->nh_gw, rta->rta_gw, 4))
+			goto err_inval;
+#else
+		goto err_inval;
+#endif
+	} else {
+		struct fib_nh *nh = fi->fib_nh;
+		if (rta->rta_oif)
+			nh->nh_oif = *rta->rta_oif;
+		if (rta->rta_gw)
+			memcpy(&nh->nh_gw, rta->rta_gw, 4);
+		nh->nh_flags = r->rtm_flags;
+#ifdef CONFIG_IP_ROUTE_MULTIPATH
+		nh->nh_weight = 1;
+#endif
+	}
+
+#ifdef CONFIG_IP_ROUTE_NAT
+	if (r->rtm_type == RTN_NAT) {
+		if (rta->rta_gw == NULL || nhs != 1 || rta->rta_oif)
+			goto err_inval;
+		memcpy(&fi->fib_nh->nh_gw, rta->rta_gw, 4);
+		goto link_it;
+	}
+#endif
+
+	if (fib_props[r->rtm_type].error) {
+		if (rta->rta_gw || rta->rta_oif || r->rtm_nhs)
+			goto err_inval;
+		goto link_it;
+	}
+
+	if (r->rtm_scope > RT_SCOPE_HOST)
+		goto err_inval;
+
+	if (r->rtm_scope == RT_SCOPE_HOST) {
+		struct fib_nh *nh = fi->fib_nh;
+
+		/* Local address is added. */
+		if (nhs != 1 || nh->nh_gw)
+			goto err_inval;
+		nh->nh_scope = RT_SCOPE_NOWHERE;
+		nh->nh_dev = dev_get_by_index(fi->fib_nh->nh_oif);
+		err = -ENODEV;
+		if (nh->nh_dev == NULL)
+			goto failure;
+	} else {
+		change_nexthops(fi) {
+			if ((err = fib_check_nh(r, fi, nh)) != 0) {
+				if (err == -EINVAL)
+					printk("Einval 2\n");
+				goto failure;
+			}
+		} endfor_nexthops(fi)
+	}
+
+	if (fi->fib_prefsrc) {
+		if (r->rtm_type != RTN_LOCAL || rta->rta_dst == NULL ||
+		    memcmp(&fi->fib_prefsrc, rta->rta_dst, 4))
+			if (inet_addr_type(fi->fib_prefsrc) != RTN_LOCAL) {
+				printk("Einval 3\n");
+				goto err_inval;
+			}
+	}
+
+link_it:
+	if ((ofi = fib_find_info(fi)) != NULL) {
+		kfree(fi);
+		ofi->fib_refcnt++;
+		return ofi;
+	}
+
+	fi->fib_refcnt++;
+	fi->fib_next = fib_info_list;
+	fi->fib_prev = NULL;
+	if (fib_info_list)
+		fib_info_list->fib_prev = fi;
+	fib_info_list = fi;
+	return fi;
+
+err_inval:
+	err = -EINVAL;
+
+failure:
+        *errp = err;
+        if (fi)
+		kfree(fi);
+	return NULL;
+}
+
+int 
+fib_semantic_match(int type, struct fib_info *fi, const struct rt_key *key, struct fib_result *res)
+{
+	int err = fib_props[type].error;
+
+	if (err == 0) {
+		if (fi->fib_flags&RTNH_F_DEAD)
+			return 1;
+
+		res->fi = fi;
+
+		switch (type) {
+#ifdef CONFIG_IP_ROUTE_NAT
+		case RTN_NAT:
+			FIB_RES_RESET(*res);
+			return 0;
+#endif
+		case RTN_UNICAST:
+		case RTN_LOCAL:
+		case RTN_BROADCAST:
+		case RTN_ANYCAST:
+		case RTN_MULTICAST:
+			for_nexthops(fi) {
+				if (nh->nh_flags&RTNH_F_DEAD)
+					continue;
+				if (!key->oif || key->oif == nh->nh_oif)
+					break;
+			}
+#ifdef CONFIG_IP_ROUTE_MULTIPATH
+			if (nhsel < fi->fib_nhs) {
+				res->nh_sel = nhsel;
+				return 0;
+			}
+#else
+			if (nhsel < 1)
+				return 0;
+#endif
+			endfor_nexthops(fi);
+			return 1;
+		default:
+			printk(KERN_DEBUG "impossible 102\n");
+			return -EINVAL;
+		}
+	}
+	return err;
+}
+
+/* Find appropriate source address to this destination */
+
+u32 __fib_res_prefsrc(struct fib_result *res)
+{
+	return inet_select_addr(FIB_RES_DEV(*res), FIB_RES_GW(*res), res->scope);
+}
+
+#ifdef CONFIG_RTNETLINK
+
+int
+fib_dump_info(struct sk_buff *skb, pid_t pid, u32 seq, int event,
+	      u8 tb_id, u8 type, u8 scope, void *dst, int dst_len, u8 tos,
+	      struct fib_info *fi)
+{
+	struct rtmsg *rtm;
+	struct nlmsghdr  *nlh;
+	unsigned char	 *b = skb->tail;
+	unsigned char 	 *o;
+
+	nlh = NLMSG_PUT(skb, pid, seq, event, sizeof(*rtm));
+	rtm = NLMSG_DATA(nlh);
+	rtm->rtm_family = AF_INET;
+	rtm->rtm_dst_len = dst_len;
+	rtm->rtm_src_len = 0;
+	rtm->rtm_tos = tos;
+	rtm->rtm_table = tb_id;
+	rtm->rtm_type = type;
+	rtm->rtm_flags = fi->fib_flags;
+	rtm->rtm_scope = scope;
+	rtm->rtm_nhs = 0;
+
+	o = skb->tail;
+	if (rtm->rtm_dst_len)
+		RTA_PUT(skb, RTA_DST, 4, dst);
+	rtm->rtm_protocol = fi->fib_protocol;
+	if (fi->fib_mtu)
+		RTA_PUT(skb, RTA_MTU, sizeof(unsigned), &fi->fib_mtu);
+	if (fi->fib_window)
+		RTA_PUT(skb, RTA_WINDOW, sizeof(unsigned), &fi->fib_window);
+	if (fi->fib_rtt)
+		RTA_PUT(skb, RTA_RTT, sizeof(unsigned), &fi->fib_rtt);
+	if (fi->fib_prefsrc)
+		RTA_PUT(skb, RTA_PREFSRC, 4, &fi->fib_prefsrc);
+	if (fi->fib_nhs == 1) {
+		if (fi->fib_nh->nh_gw)
+			RTA_PUT(skb, RTA_GATEWAY, 4, &fi->fib_nh->nh_gw);
+		if (fi->fib_nh->nh_oif)
+			RTA_PUT(skb, RTA_OIF, sizeof(int), &fi->fib_nh->nh_oif);
+	}
+	rtm->rtm_optlen = skb->tail - o;
+#ifdef CONFIG_IP_ROUTE_MULTIPATH
+	if (fi->fib_nhs > 1) {
+		struct rtnexthop *nhp;
+		for_nexthops(fi) {
+			if (skb_tailroom(skb) < RTA_ALIGN(RTA_ALIGN(sizeof(*nhp)) + 4))
+				goto rtattr_failure;
+			nhp = (struct rtnexthop*)skb_put(skb, RTA_ALIGN(sizeof(*nhp)));
+			nhp->rtnh_flags = nh->nh_flags & 0xFF;
+			nhp->rtnh_hops = nh->nh_weight-1;
+			nhp->rtnh_ifindex = nh->nh_oif;
+			if (nh->nh_gw)
+				RTA_PUT(skb, RTA_GATEWAY, 4, &nh->nh_gw);
+			nhp->rtnh_len = skb->tail - (unsigned char*)nhp;
+			rtm->rtm_nhs++;
+		} endfor_nexthops(fi);
+	}
+#endif
+	nlh->nlmsg_len = skb->tail - b;
+	return skb->len;
+
+nlmsg_failure:
+rtattr_failure:
+	skb_put(skb, b - skb->tail);
+	return -1;
+}
+
+#endif /* CONFIG_RTNETLINK */
+
+#ifndef CONFIG_IP_NOSIOCRT
+
+int
+fib_convert_rtentry(int cmd, struct nlmsghdr *nl, struct rtmsg *rtm,
+		    struct kern_rta *rta, struct rtentry *r)
+{
+	int    plen;
+	u32    *ptr;
+
+	memset(rtm, 0, sizeof(*rtm));
+	memset(rta, 0, sizeof(*rta));
+
+	if (r->rt_dst.sa_family != AF_INET)
+		return -EAFNOSUPPORT;
+
+	/* Check mask for validity:
+	   a) it must be contiguous.
+	   b) destination must have all host bits clear.
+	   c) if application forgot to set correct family (AF_INET),
+	      reject request unless it is absolutely clear i.e.
+	      both family and mask are zero.
+	 */
+	plen = 32;
+	ptr = &((struct sockaddr_in*)&r->rt_dst)->sin_addr.s_addr;
+	if (!(r->rt_flags&RTF_HOST)) {
+		u32 mask = ((struct sockaddr_in*)&r->rt_genmask)->sin_addr.s_addr;
+		if (r->rt_genmask.sa_family != AF_INET) {
+			if (mask || r->rt_genmask.sa_family)
+				return -EAFNOSUPPORT;
+		}
+		if (bad_mask(mask, *ptr))
+			return -EINVAL;
+		plen = inet_mask_len(mask);
+	}
+
+	nl->nlmsg_flags = NLM_F_REQUEST;
+	nl->nlmsg_pid = 0;
+	nl->nlmsg_seq = 0;
+	nl->nlmsg_len = NLMSG_LENGTH(sizeof(*rtm));
+	if (cmd == SIOCDELRT) {
+		nl->nlmsg_type = RTM_DELROUTE;
+		nl->nlmsg_flags = 0;
+	} else {
+		nl->nlmsg_type = RTM_NEWROUTE;
+		nl->nlmsg_flags = NLM_F_CREATE;
+		rtm->rtm_protocol = RTPROT_BOOT;
+		if (plen != 0)
+			nl->nlmsg_flags |= NLM_F_REPLACE;
+	}
+
+	rtm->rtm_dst_len = plen;
+	rta->rta_dst = ptr;
+
+	if (r->rt_flags&RTF_REJECT) {
+		rtm->rtm_scope = RT_SCOPE_HOST;
+		rtm->rtm_type = RTN_UNREACHABLE;
+		return 0;
+	}
+	rtm->rtm_scope = RT_SCOPE_LINK;
+	rtm->rtm_type = RTN_UNICAST;
+
+	if (r->rt_dev) {
+#ifdef CONFIG_IP_ALIAS
+		char *colon;
+#endif
+		struct device *dev;
+		char   devname[IFNAMSIZ];
+
+		if (copy_from_user(devname, r->rt_dev, 15))
+			return -EFAULT;
+		devname[IFNAMSIZ-1] = 0;
+#ifdef CONFIG_IP_ALIAS
+		colon = strchr(devname, ':');
+		if (colon)
+			*colon = 0;
+#endif
+		dev = dev_get(devname);
+		if (!dev)
+			return -ENODEV;
+		rta->rta_oif = &dev->ifindex;
+#ifdef CONFIG_IP_ALIAS
+		if (colon) {
+			struct in_ifaddr *ifa;
+			struct in_device *in_dev = dev->ip_ptr;
+			if (!in_dev)
+				return -ENODEV;
+			*colon = ':';
+			for (ifa = in_dev->ifa_list; ifa; ifa = ifa->ifa_next)
+				if (strcmp(ifa->ifa_label, devname) == 0)
+					break;
+			if (ifa == NULL)
+				return -ENODEV;
+			rta->rta_prefsrc = &ifa->ifa_local;
+		}
+#endif
+	}
+
+	ptr = &((struct sockaddr_in*)&r->rt_gateway)->sin_addr.s_addr;
+	if (r->rt_gateway.sa_family == AF_INET && *ptr) {
+		rta->rta_gw = ptr;
+		if (r->rt_flags&RTF_GATEWAY)
+			rtm->rtm_scope = RT_SCOPE_UNIVERSE;
+	}
+
+	if (cmd == SIOCDELRT)
+		return 0;
+
+	if (r->rt_flags&RTF_GATEWAY && rta->rta_gw == NULL)
+		return -EINVAL;
+
+	/* Ugly conversion from rtentry types to unsigned */
+
+	if (r->rt_flags&RTF_IRTT) {
+		rta->rta_rtt = (unsigned*)&r->rt_pad3;
+		*rta->rta_rtt = r->rt_irtt;
+	}
+	if (r->rt_flags&RTF_WINDOW) {
+		rta->rta_window = (unsigned*)&r->rt_window;
+		if (sizeof(*rta->rta_window) != sizeof(r->rt_window))
+			*rta->rta_window = r->rt_window;
+	}
+	if (r->rt_flags&RTF_MTU) {
+		rta->rta_mtu = (unsigned*)&r->rt_mtu;
+		if (sizeof(*rta->rta_mtu) != sizeof(r->rt_mtu))
+			*rta->rta_mtu = r->rt_mtu;
+	}
+	return 0;
+}
+
+#endif
+
+/*
+   Update FIB if:
+   - local address disappeared -> we must delete all the entries
+     referring to it.
+   - device went down -> we must shutdown all nexthops going via it.
+ */
+
+int fib_sync_down(u32 local, struct device *dev)
+{
+	int ret = 0;
+
+	for_fib_info() {
+		if (local && fi->fib_prefsrc == local) {
+			fi->fib_flags |= RTNH_F_DEAD;
+			ret++;
+		} else if (dev && fi->fib_nhs) {
+			int dead = 0;
+
+			change_nexthops(fi) {
+				if (nh->nh_flags&RTNH_F_DEAD)
+					dead++;
+				else if (nh->nh_dev == dev &&
+					 nh->nh_scope != RT_SCOPE_NOWHERE) {
+					nh->nh_flags |= RTNH_F_DEAD;
+#ifdef CONFIG_IP_ROUTE_MULTIPATH
+					fi->fib_power -= nh->nh_power;
+					nh->nh_power = 0;
+#endif
+					dead++;
+				}
+			} endfor_nexthops(fi)
+			if (dead == fi->fib_nhs) {
+				fi->fib_flags |= RTNH_F_DEAD;
+				ret++;
+			}
+		}
+	} endfor_fib_info();
+	return ret;
+}
+
+#ifdef CONFIG_IP_ROUTE_MULTIPATH
+
+/*
+   Dead device goes up. We wake up dead nexthops.
+   It takes sense only on multipath routes.
+ */
+
+int fib_sync_up(struct device *dev)
+{
+	int ret = 0;
+
+	if (!(dev->flags&IFF_UP))
+		return 0;
+
+	for_fib_info() {
+		int alive = 0;
+
+		change_nexthops(fi) {
+			if (!(nh->nh_flags&RTNH_F_DEAD)) {
+				alive++;
+				continue;
+			}
+			if (nh->nh_dev == NULL || !(nh->nh_dev->flags&IFF_UP))
+				continue;
+			if (nh->nh_dev != dev || dev->ip_ptr == NULL)
+				continue;
+			alive++;
+			nh->nh_power = 0;
+			nh->nh_flags &= ~RTNH_F_DEAD;
+		} endfor_nexthops(fi)
+
+		if (alive == fi->fib_nhs) {
+			fi->fib_flags &= ~RTNH_F_DEAD;
+			ret++;
+		}
+	} endfor_fib_info();
+	return ret;
+}
+
+/*
+   The algorithm is suboptimal, but it provides really
+   fair weighted route distribution.
+ */
+
+void fib_select_multipath(const struct rt_key *key, struct fib_result *res)
+{
+	struct fib_info *fi = res->fi;
+	int w;
+
+	if (fi->fib_power <= 0) {
+		int power = 0;
+		change_nexthops(fi) {
+			if (!(nh->nh_flags&RTNH_F_DEAD)) {
+				power += nh->nh_weight;
+				nh->nh_power = nh->nh_weight;
+			}
+		} endfor_nexthops(fi);
+		fi->fib_power = power;
+#if 1
+		if (power <= 0) {
+			printk(KERN_CRIT "impossible 777\n");
+			return;
+		}
+#endif
+	}
+
+
+	/* w should be random number [0..fi->fib_power-1],
+	   it is pretty bad approximation.
+	 */
+
+	w = jiffies % fi->fib_power;
+
+	change_nexthops(fi) {
+		if (!(nh->nh_flags&RTNH_F_DEAD) && nh->nh_power) {
+			if ((w -= nh->nh_power) <= 0) {
+				nh->nh_power--;
+				fi->fib_power--;
+				res->nh_sel = nhsel;
+				return;
+			}
+		}
+	} endfor_nexthops(fi);
+
+#if 1
+	printk(KERN_CRIT "impossible 888\n");
+#endif
+	return;
+}
+#endif
+
+
+#ifdef CONFIG_PROC_FS
+
+static unsigned fib_flag_trans(int type, int dead, u32 mask, struct fib_info *fi)
+{
+	static unsigned type2flags[RTN_MAX+1] = {
+		0, 0, 0, 0, 0, 0, 0, RTF_REJECT, RTF_REJECT, 0, 0, 0
+	};
+	unsigned flags = type2flags[type];
+
+	if (fi && fi->fib_nh->nh_gw)
+		flags |= RTF_GATEWAY;
+	if (mask == 0xFFFFFFFF)
+		flags |= RTF_HOST;
+	if (!dead)
+		flags |= RTF_UP;
+	return flags;
+}
+
+void fib_node_get_info(int type, int dead, struct fib_info *fi, u32 prefix, u32 mask, char *buffer)
+{
+	int len;
+	unsigned flags = fib_flag_trans(type, dead, mask, fi);
+
+	if (fi) {
+		len = sprintf(buffer, "%s\t%08X\t%08X\t%04X\t%d\t%u\t%d\t%08X\t%d\t%u\t%u",
+			      fi->fib_dev ? fi->fib_dev->name : "*", prefix,
+			      fi->fib_nh->nh_gw, flags, 0, 0, 0,
+			      mask, fi->fib_mtu, fi->fib_window, fi->fib_rtt);
+	} else {
+		len = sprintf(buffer, "*\t%08X\t%08X\t%04X\t%d\t%u\t%d\t%08X\t%d\t%u\t%u",
+			      prefix, 0,
+			      flags, 0, 0, 0,
+			      mask, 0, 0, 0);
+	}
+	memset(buffer+len, 0, 127-len);
+	buffer[127] = '\n';
+}
+
+#endif
diff --git a/net/ipv4/ip_alias.c b/net/ipv4/ip_alias.c
deleted file mode 100644
index e69de29bb..000000000
--- a/net/ipv4/ip_alias.c
+++ /dev/null
diff --git a/net/ipv4/ip_gre.c b/net/ipv4/ip_gre.c
new file mode 100644
index 000000000..dbd62e27e
--- /dev/null
+++ b/net/ipv4/ip_gre.c
@@ -0,0 +1,1191 @@
+/*
+ *	Linux NET3:	GRE over IP protocol decoder. 
+ *
+ *	Authors: Alexey Kuznetsov (kuznet@ms2.inr.ac.ru)
+ *
+ *	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.
+ *
+ */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <linux/types.h>
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <asm/uaccess.h>
+#include <linux/skbuff.h>
+#include <linux/netdevice.h>
+#include <linux/in.h>
+#include <linux/tcp.h>
+#include <linux/udp.h>
+#include <linux/if_arp.h>
+#include <linux/mroute.h>
+#include <linux/init.h>
+#include <linux/in6.h>
+#include <linux/inetdevice.h>
+#include <linux/igmp.h>
+
+#include <net/sock.h>
+#include <net/ip.h>
+#include <net/icmp.h>
+#include <net/protocol.h>
+#include <net/ipip.h>
+#include <net/arp.h>
+#include <net/checksum.h>
+
+#ifdef CONFIG_IPV6
+#include <net/ipv6.h>
+#include <net/ip6_fib.h>
+#include <net/ip6_route.h>
+#endif
+
+/*
+   Problems & solutions
+   --------------------
+
+   1. The most important issue is detecting local dead loops.
+   They would cause complete host lockup in transmit, which
+   would be "resolved" by stack overflow or, if queueing is enabled,
+   with infinite looping in net_bh.
+
+   We cannot track such dead loops during route installation,
+   it is infeasible task. The most general solutions would be
+   to keep skb->encapsulation counter (sort of local ttl),
+   and silently drop packet when it expires. It is the best
+   solution, but it supposes maintaing new variable in ALL
+   skb, even if no tunneling is used.
+
+   Current solution: t->recursion lock breaks dead loops. It looks 
+   like dev->tbusy flag, but I preferred new variable, because
+   the semantics is different. One day, when hard_start_xmit
+   will be multithreaded we will have to use skb->encapsulation.
+
+
+
+   2. Networking dead loops would not kill routers, but would really
+   kill network. IP hop limit plays role of "t->recursion" in this case,
+   if we copy it from packet being encapsulated to upper header.
+   It is very good solution, but it introduces two problems:
+
+   - Routing protocols, using packets with ttl=1 (OSPF, RIP2),
+     do not work over tunnels.
+   - traceroute does not work. I planned to relay ICMP from tunnel,
+     so that this problem would be solved and traceroute output
+     would even more informative. This idea appeared to be wrong:
+     only Linux complies to rfc1812 now (yes, guys, Linux is the only
+     true router now :-)), all routers (at least, in neighbourhood of mine)
+     return only 8 bytes of payload. It is the end.
+
+   Hence, if we want that OSPF worked or traceroute said something reasonable,
+   we should search for another solution.
+
+   One of them is to parse packet trying to detect inner encapsulation
+   made by our node. It is difficult or even impossible, especially,
+   taking into account fragmentation. TO be short, tt is not solution at all.
+
+   Current solution: The solution was UNEXPECTEDLY SIMPLE.
+   We force DF flag on tunnels with preconfigured hop limit,
+   that is ALL. :-) Well, it does not remove the problem completely,
+   but exponential growth of network traffic is changed to linear
+   (branches, that exceed pmtu are pruned) and tunnel mtu
+   fastly degrades to value <68, where looping stops.
+   Yes, it is not good if there exists a router in the loop,
+   which does not force DF, even when encapsulating packets have DF set.
+   But it is not our problem! Nobody could accuse us, we made
+   all that we could make. Even if it is your gated who injected
+   fatal route to network, even if it were you who configured
+   fatal static route: you are innocent. :-)
+
+
+
+   3. Really, ipv4/ipip.c, ipv4/ip_gre.c and ipv6/sit.c contain
+   practically identical code. It would be good to glue them
+   together, but it is not very evident, how to make them modular.
+   sit is integral part of IPv6, ipip and gre are naturally modular.
+   We could extract common parts (hash table, ioctl etc)
+   to a separate module (ip_tunnel.c).
+
+   Alexey Kuznetsov.
+ */
+
+static int ipgre_tunnel_init(struct device *dev);
+
+/* Fallback tunnel: no source, no destination, no key, no options */
+
+static int ipgre_fb_tunnel_init(struct device *dev);
+
+static struct device ipgre_fb_tunnel_dev = {
+	NULL, 0x0, 0x0, 0x0, 0x0, 0, 0, 0, 0, 0, NULL, ipgre_fb_tunnel_init,
+};
+
+static struct ip_tunnel ipgre_fb_tunnel = {
+	NULL, &ipgre_fb_tunnel_dev, {0, }, 0, 0, 0, 0, 0, 0, 0, {"gre0", }
+};
+
+/* Tunnel hash table */
+
+/*
+   4 hash tables:
+
+   3: (remote,local)
+   2: (remote,*)
+   1: (*,local)
+   0: (*,*)
+
+   We require exact key match i.e. if a key is present in packet
+   it will match only tunnel with the same key; if it is not present,
+   it will match only keyless tunnel.
+
+   All keysless packets, if not matched configured keyless tunnels
+   will match fallback tunnel.
+ */
+
+#define HASH_SIZE  16
+#define HASH(addr) ((addr^(addr>>4))&0xF)
+
+static struct ip_tunnel *tunnels[4][HASH_SIZE];
+
+#define tunnels_r_l	(tunnels[3])
+#define tunnels_r	(tunnels[2])
+#define tunnels_l	(tunnels[1])
+#define tunnels_wc	(tunnels[0])
+
+/* Given src, dst and key, find approriate for input tunnel. */
+
+static struct ip_tunnel * ipgre_tunnel_lookup(u32 remote, u32 local, u32 key)
+{
+	unsigned h0 = HASH(remote);
+	unsigned h1 = HASH(key);
+	struct ip_tunnel *t;
+
+	for (t = tunnels_r_l[h0^h1]; t; t = t->next) {
+		if (local == t->parms.iph.saddr && remote == t->parms.iph.daddr) {
+			if (t->parms.i_key == key && (t->dev->flags&IFF_UP))
+				return t;
+		}
+	}
+	for (t = tunnels_r[h0^h1]; t; t = t->next) {
+		if (remote == t->parms.iph.daddr) {
+			if (t->parms.i_key == key && (t->dev->flags&IFF_UP))
+				return t;
+		}
+	}
+	for (t = tunnels_l[h1]; t; t = t->next) {
+		if (local == t->parms.iph.saddr ||
+		     (local == t->parms.iph.daddr && MULTICAST(local))) {
+			if (t->parms.i_key == key && (t->dev->flags&IFF_UP))
+				return t;
+		}
+	}
+	for (t = tunnels_wc[h1]; t; t = t->next) {
+		if (t->parms.i_key == key && (t->dev->flags&IFF_UP))
+			return t;
+	}
+	if (ipgre_fb_tunnel_dev.flags&IFF_UP)
+		return &ipgre_fb_tunnel;
+	return NULL;
+}
+
+static struct ip_tunnel * ipgre_tunnel_locate(struct ip_tunnel_parm *parms, int create)
+{
+	u32 remote = parms->iph.daddr;
+	u32 local = parms->iph.saddr;
+	u32 key = parms->i_key;
+	struct ip_tunnel *t, **tp, *nt;
+	struct device *dev;
+	unsigned h = HASH(key);
+	int prio = 0;
+
+	if (local)
+		prio |= 1;
+	if (remote && !MULTICAST(remote)) {
+		prio |= 2;
+		h ^= HASH(remote);
+	}
+	for (tp = &tunnels[prio][h]; (t = *tp) != NULL; tp = &t->next) {
+		if (local == t->parms.iph.saddr && remote == t->parms.iph.daddr) {
+			if (key == t->parms.i_key)
+				return t;
+		}
+	}
+	if (!create)
+		return NULL;
+
+	MOD_INC_USE_COUNT;
+	dev = kmalloc(sizeof(*dev) + sizeof(*t), GFP_KERNEL);
+	if (dev == NULL) {
+		MOD_DEC_USE_COUNT;
+		return NULL;
+	}
+	memset(dev, 0, sizeof(*dev) + sizeof(*t));
+	dev->priv = (void*)(dev+1);
+	nt = (struct ip_tunnel*)dev->priv;
+	nt->dev = dev;
+	dev->name = nt->parms.name;
+	dev->init = ipgre_tunnel_init;
+	memcpy(&nt->parms, parms, sizeof(*parms));
+	if (dev->name[0] == 0) {
+		int i;
+		for (i=1; i<100; i++) {
+			sprintf(dev->name, "gre%d", i);
+			if (dev_get(dev->name) == NULL)
+				break;
+		}
+		if (i==100)
+			goto failed;
+		memcpy(parms->name, dev->name, IFNAMSIZ);
+	}
+	if (register_netdevice(dev) < 0)
+		goto failed;
+
+	start_bh_atomic();
+	nt->next = t;
+	*tp = nt;
+	end_bh_atomic();
+	/* Do not decrement MOD_USE_COUNT here. */
+	return nt;
+
+failed:
+	kfree(dev);
+	MOD_DEC_USE_COUNT;
+	return NULL;
+}
+
+static void ipgre_tunnel_destroy(struct device *dev)
+{
+	struct ip_tunnel *t, **tp;
+	struct ip_tunnel *t0 = (struct ip_tunnel*)dev->priv;
+	u32 remote = t0->parms.iph.daddr;
+	u32 local = t0->parms.iph.saddr;
+	unsigned h = HASH(t0->parms.i_key);
+	int prio = 0;
+
+	if (local)
+		prio |= 1;
+	if (remote && !MULTICAST(remote)) {
+		prio |= 2;
+		h ^= HASH(remote);
+	}
+	for (tp = &tunnels[prio][h]; (t = *tp) != NULL; tp = &t->next) {
+		if (t == t0) {
+			*tp = t->next;
+			if (dev != &ipgre_fb_tunnel_dev) {
+				kfree(dev);
+				MOD_DEC_USE_COUNT;
+			}
+			break;
+		}
+	}
+}
+
+
+void ipgre_err(struct sk_buff *skb, unsigned char *dp, int len)
+{
+#ifndef I_WISH_WORLD_WERE_PERFECT
+
+/* It is not :-( All the routers (except for Linux) return only
+   8 bytes of packet payload. It means, that precise relaying of
+   ICMP in the real Internet is absolutely infeasible.
+
+   Moreover, Cisco "wise men" put GRE key to the third word
+   in GRE header. It makes impossible maintaining even soft state for keyed
+   GRE tunnels with enabled checksum. Tell them "thank you".
+
+   Well, I wonder, rfc1812 was written by Cisco employee,
+   what the hell these idiots break standrads established
+   by themself???
+ */
+
+	struct iphdr *iph = (struct iphdr*)dp;
+	u16	     *p = (u16*)(dp+(iph->ihl<<2));
+	int grehlen = (iph->ihl<<2) + 4;
+	int type = skb->h.icmph->type;
+	int code = skb->h.icmph->code;
+	struct ip_tunnel *t;
+	u16 flags;
+
+	flags = p[0];
+	if (flags&(GRE_CSUM|GRE_KEY|GRE_SEQ|GRE_ROUTING|GRE_VERSION)) {
+		if (flags&(GRE_VERSION|GRE_ROUTING))
+			return;
+		if (flags&GRE_KEY) {
+			grehlen += 4;
+			if (flags&GRE_CSUM)
+				grehlen += 4;
+		}
+	}
+
+	/* If only 8 bytes returned, keyed message will be dropped here */
+	if (len < grehlen)
+		return;
+
+	switch (type) {
+	default:
+	case ICMP_PARAMETERPROB:
+		return;
+
+	case ICMP_DEST_UNREACH:
+		switch (code) {
+		case ICMP_SR_FAILED:
+		case ICMP_PORT_UNREACH:
+			/* Impossible event. */
+			return;
+		case ICMP_FRAG_NEEDED:
+			/* Soft state for pmtu is maintained by IP core. */
+			return;
+		default:
+			/* All others are translated to HOST_UNREACH.
+			   rfc2003 contains "deep thoughts" about NET_UNREACH,
+			   I believe they are just ether pollution. --ANK
+			 */
+			break;
+		}
+		break;
+	case ICMP_TIME_EXCEEDED:
+		if (code != ICMP_EXC_TTL)
+			return;
+		break;
+	}
+
+	t = ipgre_tunnel_lookup(iph->daddr, iph->saddr, (flags&GRE_KEY) ? *(((u32*)p) + (grehlen>>2) - 1) : 0);
+	if (t == NULL || t->parms.iph.daddr == 0 || MULTICAST(t->parms.iph.daddr))
+		return;
+
+	if (t->parms.iph.ttl == 0 && type == ICMP_TIME_EXCEEDED)
+		return;
+
+	if (jiffies - t->err_time < IPTUNNEL_ERR_TIMEO)
+		t->err_count++;
+	else
+		t->err_count = 1;
+	t->err_time = jiffies;
+	return;
+#else
+	struct iphdr *iph = (struct iphdr*)dp;
+	struct iphdr *eiph;
+	u16	     *p = (u16*)(dp+(iph->ihl<<2));
+	int type = skb->h.icmph->type;
+	int code = skb->h.icmph->code;
+	int rel_type = 0;
+	int rel_code = 0;
+	int rel_info = 0;
+	u16 flags;
+	int grehlen = (iph->ihl<<2) + 4;
+	struct sk_buff *skb2;
+	struct rtable *rt;
+
+	if (p[1] != __constant_htons(ETH_P_IP))
+		return;
+
+	flags = p[0];
+	if (flags&(GRE_CSUM|GRE_KEY|GRE_SEQ|GRE_ROUTING|GRE_VERSION)) {
+		if (flags&(GRE_VERSION|GRE_ROUTING))
+			return;
+		if (flags&GRE_CSUM)
+			grehlen += 4;
+		if (flags&GRE_KEY)
+			grehlen += 4;
+		if (flags&GRE_SEQ)
+			grehlen += 4;
+	}
+	if (len < grehlen + sizeof(struct iphdr))
+		return;
+	eiph = (struct iphdr*)(dp + grehlen);
+
+	switch (type) {
+	default:
+		return;
+	case ICMP_PARAMETERPROB:
+		if (skb->h.icmph->un.gateway < (iph->ihl<<2))
+			return;
+
+		/* So... This guy found something strange INSIDE encapsulated
+		   packet. Well, he is fool, but what can we do ?
+		 */
+		rel_type = ICMP_PARAMETERPROB;
+		rel_info = skb->h.icmph->un.gateway - grehlen;
+		break;
+
+	case ICMP_DEST_UNREACH:
+		switch (code) {
+		case ICMP_SR_FAILED:
+		case ICMP_PORT_UNREACH:
+			/* Impossible event. */
+			return;
+		case ICMP_FRAG_NEEDED:
+			/* And it is the only really necesary thing :-) */
+			rel_info = ntohs(skb->h.icmph->un.frag.mtu);
+			if (rel_info < grehlen+68)
+				return;
+			rel_info -= grehlen;
+			/* BSD 4.2 MORE DOES NOT EXIST IN NATURE. */
+			if (rel_info > ntohs(eiph->tot_len))
+				return;
+			break;
+		default:
+			/* All others are translated to HOST_UNREACH.
+			   rfc2003 contains "deep thoughts" about NET_UNREACH,
+			   I believe, it is just ether pollution. --ANK
+			 */
+			rel_type = ICMP_DEST_UNREACH;
+			rel_code = ICMP_HOST_UNREACH;
+			break;
+		}
+		break;
+	case ICMP_TIME_EXCEEDED:
+		if (code != ICMP_EXC_TTL)
+			return;
+		break;
+	}
+
+	/* Prepare fake skb to feed it to icmp_send */
+	skb2 = skb_clone(skb, GFP_ATOMIC);
+	if (skb2 == NULL)
+		return;
+	dst_release(skb2->dst);
+	skb2->dst = NULL;
+	skb_pull(skb2, skb->data - (u8*)eiph);
+	skb2->nh.raw = skb2->data;
+
+	/* Try to guess incoming interface */
+	if (ip_route_output(&rt, eiph->saddr, 0, RT_TOS(eiph->tos), 0)) {
+		kfree_skb(skb2, FREE_WRITE);
+		return;
+	}
+	skb2->dev = rt->u.dst.dev;
+
+	/* route "incoming" packet */
+	if (rt->rt_flags&RTCF_LOCAL) {
+		ip_rt_put(rt);
+		rt = NULL;
+		if (ip_route_output(&rt, eiph->daddr, eiph->saddr, eiph->tos, 0) ||
+		    rt->u.dst.dev->type != ARPHRD_IPGRE) {
+			ip_rt_put(rt);
+			kfree_skb(skb2, FREE_WRITE);
+			return;
+		}
+	} else {
+		ip_rt_put(rt);
+		if (ip_route_input(skb2, eiph->daddr, eiph->saddr, eiph->tos, skb2->dev) ||
+		    skb2->dst->dev->type != ARPHRD_IPGRE) {
+			kfree_skb(skb2, FREE_WRITE);
+			return;
+		}
+	}
+
+	/* change mtu on this route */
+	if (type == ICMP_DEST_UNREACH && code == ICMP_FRAG_NEEDED) {
+		if (rel_info > skb2->dst->pmtu) {
+			kfree_skb(skb2, FREE_WRITE);
+			return;
+		}
+		skb2->dst->pmtu = rel_info;
+		rel_info = htonl(rel_info);
+	} else if (type == ICMP_TIME_EXCEEDED) {
+		struct ip_tunnel *t = (struct ip_tunnel*)skb2->dev->priv;
+		if (t->parms.iph.ttl) {
+			rel_type = ICMP_DEST_UNREACH;
+			rel_code = ICMP_HOST_UNREACH;
+		}
+	}
+
+	icmp_send(skb2, rel_type, rel_code, rel_info);
+	kfree_skb(skb2, FREE_WRITE);
+#endif
+}
+
+int ipgre_rcv(struct sk_buff *skb, unsigned short len)
+{
+	struct iphdr *iph = skb->nh.iph;
+	u8     *h = skb->h.raw;
+	u16    flags = *(u16*)h;
+	u16    csum = 0;
+	u32    key = 0;
+	u32    seqno = 0;
+	struct ip_tunnel *tunnel;
+	int    offset = 4;
+
+	if (flags&(GRE_CSUM|GRE_KEY|GRE_ROUTING|GRE_SEQ|GRE_VERSION)) {
+		/* - Version must be 0.
+		   - We do not support routing headers.
+		 */
+		if (flags&(GRE_VERSION|GRE_ROUTING))
+			goto drop;
+
+		if (flags&GRE_CSUM) {
+			csum = ip_compute_csum(h, len);
+			offset += 4;
+		}
+		if (flags&GRE_KEY) {
+			key = *(u32*)(h + offset);
+			offset += 4;
+		}
+		if (flags&GRE_SEQ) {
+			seqno = ntohl(*(u32*)(h + offset));
+			offset += 4;
+		}
+	}
+
+	if ((tunnel = ipgre_tunnel_lookup(iph->saddr, iph->daddr, key)) != NULL) {
+		skb->mac.raw = skb->nh.raw;
+		skb->nh.raw = skb_pull(skb, h + offset - skb->data);
+		memset(&(IPCB(skb)->opt), 0, sizeof(struct ip_options));
+		skb->ip_summed = 0;
+		skb->protocol = *(u16*)(h + 2);
+		skb->pkt_type = PACKET_HOST;
+#ifdef CONFIG_NET_IPGRE_BROADCAST
+		if (MULTICAST(iph->daddr)) {
+			/* Looped back packet, drop it! */
+			if (((struct rtable*)skb->dst)->key.iif == 0)
+				goto drop;
+			tunnel->stat.multicast++;
+			skb->pkt_type = PACKET_BROADCAST;
+		}
+#endif
+
+		if (((flags&GRE_CSUM) && csum) ||
+		    (!(flags&GRE_CSUM) && tunnel->parms.i_flags&GRE_CSUM)) {
+			tunnel->stat.rx_crc_errors++;
+			tunnel->stat.rx_errors++;
+			goto drop;
+		}
+		if (tunnel->parms.i_flags&GRE_SEQ) {
+			if (!(flags&GRE_SEQ) ||
+			    (tunnel->i_seqno && seqno - tunnel->i_seqno < 0)) {
+				tunnel->stat.rx_fifo_errors++;
+				tunnel->stat.rx_errors++;
+				goto drop;
+			}
+			tunnel->i_seqno = seqno + 1;
+		}
+		tunnel->stat.rx_packets++;
+		tunnel->stat.rx_bytes += skb->len;
+		skb->dev = tunnel->dev;
+		dst_release(skb->dst);
+		skb->dst = NULL;
+		netif_rx(skb);
+		return(0);
+	}
+	icmp_send(skb, ICMP_DEST_UNREACH, ICMP_PROT_UNREACH, 0);
+
+drop:
+	kfree_skb(skb, FREE_READ);
+	return(0);
+}
+
+static int ipgre_tunnel_xmit(struct sk_buff *skb, struct device *dev)
+{
+	struct ip_tunnel *tunnel = (struct ip_tunnel*)dev->priv;
+	struct net_device_stats *stats = &tunnel->stat;
+	struct iphdr  *old_iph = skb->nh.iph;
+	struct iphdr  *tiph;
+	u8     tos;
+	u16    df;
+	struct rtable *rt;     			/* Route to the other host */
+	struct device *tdev;			/* Device to other host */
+	struct iphdr  *iph;			/* Our new IP header */
+	int    max_headroom;			/* The extra header space needed */
+	int    gre_hlen;
+	u32    dst;
+	int    mtu;
+
+	if (tunnel->recursion++) {
+		tunnel->stat.collisions++;
+		goto tx_error;
+	}
+
+	if (dev->hard_header) {
+		gre_hlen = 0;
+		tiph = (struct iphdr*)skb->data;
+	} else {
+		gre_hlen = tunnel->hlen;
+		tiph = &tunnel->parms.iph;
+	}
+
+	if ((dst = tiph->daddr) == 0) {
+		/* NBMA tunnel */
+
+		if (skb->dst == NULL) {
+			tunnel->stat.tx_fifo_errors++;
+			goto tx_error;
+		}
+
+		if (skb->protocol == __constant_htons(ETH_P_IP)) {
+			rt = (struct rtable*)skb->dst;
+			if ((dst = rt->rt_gateway) == 0)
+				goto tx_error_icmp;
+		}
+#ifdef CONFIG_IPV6
+		else if (skb->protocol == __constant_htons(ETH_P_IPV6)) {
+			struct in6_addr *addr6;
+			int addr_type;
+			struct nd_neigh *neigh = (struct nd_neigh *) skb->dst->neighbour;
+
+			if (neigh == NULL)
+				goto tx_error;
+
+			addr6 = &neigh->ndn_addr;
+			addr_type = ipv6_addr_type(addr6);
+
+			if (addr_type == IPV6_ADDR_ANY) {
+				addr6 = &skb->nh.ipv6h->daddr;
+				addr_type = ipv6_addr_type(addr6);
+			}
+
+			if ((addr_type & IPV6_ADDR_COMPATv4) == 0)
+				goto tx_error_icmp;
+
+			dst = addr6->s6_addr32[3];
+		}
+#endif
+		else
+			goto tx_error;
+	}
+
+	tos = tiph->tos;
+	if (tos&1) {
+		if (skb->protocol == __constant_htons(ETH_P_IP))
+			tos = old_iph->tos;
+		tos &= ~1;
+	}
+
+	if (ip_route_output(&rt, dst, tiph->saddr, RT_TOS(tos), tunnel->parms.link)) {
+		tunnel->stat.tx_carrier_errors++;
+		goto tx_error;
+	}
+	tdev = rt->u.dst.dev;
+
+	if (tdev == dev) {
+		ip_rt_put(rt);
+		tunnel->stat.collisions++;
+		goto tx_error;
+	}
+
+	df = tiph->frag_off;
+	mtu = rt->u.dst.pmtu - tunnel->hlen;
+
+	if (skb->protocol == __constant_htons(ETH_P_IP)) {
+		if (skb->dst && mtu < skb->dst->pmtu && mtu >= 68)
+			skb->dst->pmtu = mtu;
+
+		df |= (old_iph->frag_off&__constant_htons(IP_DF));
+
+		if ((old_iph->frag_off&__constant_htons(IP_DF)) &&
+		    mtu < ntohs(old_iph->tot_len)) {
+			icmp_send(skb, ICMP_DEST_UNREACH, ICMP_FRAG_NEEDED, htonl(mtu));
+			ip_rt_put(rt);
+			goto tx_error;
+		}
+	}
+#ifdef CONFIG_IPV6
+	else if (skb->protocol == __constant_htons(ETH_P_IPV6)) {
+		struct rt6_info *rt6 = (struct rt6_info*)skb->dst;
+
+		if (rt6 && mtu < rt6->u.dst.pmtu && mtu >= 576) {
+			if ((tunnel->parms.iph.daddr && !MULTICAST(tunnel->parms.iph.daddr)) ||
+			    rt6->rt6i_dst.plen == 128) {
+				rt6->rt6i_flags |= RTF_MODIFIED;
+				skb->dst->pmtu = mtu;
+			}
+		}
+
+		if (mtu >= 576 && mtu < skb->len - tunnel->hlen + gre_hlen) {
+			icmpv6_send(skb, ICMPV6_PKT_TOOBIG, 0, mtu, dev);
+			ip_rt_put(rt);
+			goto tx_error;
+		}
+	}
+#endif
+
+	if (tunnel->err_count > 0) {
+		if (jiffies - tunnel->err_time < IPTUNNEL_ERR_TIMEO) {
+			tunnel->err_count--;
+
+			if (skb->protocol == __constant_htons(ETH_P_IP))
+				icmp_send(skb, ICMP_DEST_UNREACH, ICMP_HOST_UNREACH, 0);
+#ifdef CONFIG_IPV6
+			else if (skb->protocol == __constant_htons(ETH_P_IPV6))
+				icmpv6_send(skb, ICMPV6_DEST_UNREACH, ICMPV6_ADDR_UNREACH, 0, dev);
+#endif
+		} else
+			tunnel->err_count = 0;
+	}
+
+	skb->h.raw = skb->nh.raw;
+
+	max_headroom = ((tdev->hard_header_len+15)&~15)+ gre_hlen;
+
+	if (skb_headroom(skb) < max_headroom || skb_cloned(skb) || skb_shared(skb)) {
+		struct sk_buff *new_skb = skb_realloc_headroom(skb, max_headroom);
+		if (!new_skb) {
+			ip_rt_put(rt);
+  			stats->tx_dropped++;
+			dev_kfree_skb(skb, FREE_WRITE);
+			tunnel->recursion--;
+			return 0;
+		}
+		dev_kfree_skb(skb, FREE_WRITE);
+		skb = new_skb;
+	}
+
+	skb->nh.raw = skb_push(skb, gre_hlen);
+	memset(&(IPCB(skb)->opt), 0, sizeof(IPCB(skb)->opt));
+	dst_release(skb->dst);
+	skb->dst = &rt->u.dst;
+
+	/*
+	 *	Push down and install the IPIP header.
+	 */
+
+	iph 			=	skb->nh.iph;
+	iph->version		=	4;
+	iph->ihl		=	sizeof(struct iphdr) >> 2;
+	iph->frag_off		=	df;
+	iph->protocol		=	IPPROTO_GRE;
+	iph->tos		=	tos;
+	iph->daddr		=	rt->rt_dst;
+	iph->saddr		=	rt->rt_src;
+
+	if ((iph->ttl = tiph->ttl) == 0) {
+		if (skb->protocol == __constant_htons(ETH_P_IP))
+			iph->ttl = old_iph->ttl;
+#ifdef CONFIG_IPV6
+		else if (skb->protocol == __constant_htons(ETH_P_IPV6))
+			iph->ttl = ((struct ipv6hdr*)old_iph)->hop_limit;
+#endif
+		else
+			iph->ttl = ip_statistics.IpDefaultTTL;
+	}
+
+	((u16*)(iph+1))[0] = tunnel->parms.o_flags;
+	((u16*)(iph+1))[1] = skb->protocol;
+
+	if (tunnel->parms.o_flags&(GRE_KEY|GRE_CSUM|GRE_SEQ)) {
+		u32 *ptr = (u32*)(((u8*)iph) + tunnel->hlen - 4);
+
+		if (tunnel->parms.o_flags&GRE_SEQ) {
+			++tunnel->o_seqno;
+			*ptr = htonl(tunnel->o_seqno);
+			ptr--;
+		}
+		if (tunnel->parms.o_flags&GRE_KEY) {
+			*ptr = tunnel->parms.o_key;
+			ptr--;
+		}
+		if (tunnel->parms.o_flags&GRE_CSUM) {
+			*ptr = 0;
+			*(__u16*)ptr = ip_compute_csum((void*)(iph+1), skb->len - sizeof(struct iphdr));
+		}
+	}
+
+	iph->tot_len		=	htons(skb->len);
+	iph->id			=	htons(ip_id_count++);
+	ip_send_check(iph);
+
+	stats->tx_bytes += skb->len;
+	stats->tx_packets++;
+	ip_send(skb);
+	tunnel->recursion--;
+	return 0;
+
+tx_error_icmp:
+	if (skb->protocol == __constant_htons(ETH_P_IP))
+		icmp_send(skb, ICMP_DEST_UNREACH, ICMP_HOST_UNREACH, 0);
+#ifdef CONFIG_IPV6
+	else if (skb->protocol == __constant_htons(ETH_P_IPV6))
+		icmpv6_send(skb, ICMPV6_DEST_UNREACH, ICMPV6_ADDR_UNREACH, 0, dev);
+#endif
+
+tx_error:
+	stats->tx_errors++;
+	dev_kfree_skb(skb, FREE_WRITE);
+	tunnel->recursion--;
+	return 0;
+}
+
+static int
+ipgre_tunnel_ioctl (struct device *dev, struct ifreq *ifr, int cmd)
+{
+	int err = 0;
+	struct ip_tunnel_parm p;
+	struct ip_tunnel *t;
+
+	MOD_INC_USE_COUNT;
+
+	switch (cmd) {
+	case SIOCGETTUNNEL:
+		t = NULL;
+		if (dev == &ipgre_fb_tunnel_dev) {
+			if (copy_from_user(&p, ifr->ifr_ifru.ifru_data, sizeof(p))) {
+				err = -EFAULT;
+				break;
+			}
+			t = ipgre_tunnel_locate(&p, 0);
+		}
+		if (t == NULL)
+			t = (struct ip_tunnel*)dev->priv;
+		memcpy(&p, &t->parms, sizeof(p));
+		if (copy_to_user(ifr->ifr_ifru.ifru_data, &p, sizeof(p)))
+			err = -EFAULT;
+		break;
+
+	case SIOCADDTUNNEL:
+	case SIOCCHGTUNNEL:
+		err = -EFAULT;
+		if (copy_from_user(&p, ifr->ifr_ifru.ifru_data, sizeof(p)))
+			goto done;
+
+		err = -EINVAL;
+		if (p.iph.version != 4 || p.iph.protocol != IPPROTO_GRE ||
+		    p.iph.ihl != 5 || (p.iph.frag_off&__constant_htons(~IP_DF)) ||
+		    ((p.i_flags|p.o_flags)&(GRE_VERSION|GRE_ROUTING)))
+			goto done;
+		if (p.iph.ttl)
+			p.iph.frag_off |= __constant_htons(IP_DF);
+
+		if (!(p.i_flags&GRE_KEY))
+			p.i_key = 0;
+		if (!(p.o_flags&GRE_KEY))
+			p.o_key = 0;
+
+		t = ipgre_tunnel_locate(&p, cmd == SIOCADDTUNNEL);
+
+		if (t) {
+			err = 0;
+			if (cmd == SIOCCHGTUNNEL) {
+				t->parms.iph.ttl = p.iph.ttl;
+				t->parms.iph.tos = p.iph.tos;
+				t->parms.iph.frag_off = p.iph.frag_off;
+			}
+			if (copy_to_user(ifr->ifr_ifru.ifru_data, &t->parms, sizeof(p)))
+				err = -EFAULT;
+		} else
+			err = (cmd == SIOCADDTUNNEL ? -ENOBUFS : -ENOENT);
+		break;
+
+	case SIOCDELTUNNEL:
+		if (dev == &ipgre_fb_tunnel_dev) {
+			err = -EFAULT;
+			if (copy_from_user(&p, ifr->ifr_ifru.ifru_data, sizeof(p)))
+				goto done;
+			err = -ENOENT;
+			if ((t = ipgre_tunnel_locate(&p, 0)) == NULL)
+				goto done;
+			err = -EPERM;
+			if (t == &ipgre_fb_tunnel)
+				goto done;
+		}
+		err = unregister_netdevice(dev);
+		break;
+
+	default:
+		err = -EINVAL;
+	}
+
+done:
+	MOD_DEC_USE_COUNT;
+	return err;
+}
+
+static struct net_device_stats *ipgre_tunnel_get_stats(struct device *dev)
+{
+	return &(((struct ip_tunnel*)dev->priv)->stat);
+}
+
+static int ipgre_tunnel_change_mtu(struct device *dev, int new_mtu)
+{
+	struct ip_tunnel *tunnel = (struct ip_tunnel*)dev->priv;
+	if (new_mtu < 68 || new_mtu > 0xFFF8 - tunnel->hlen)
+		return -EINVAL;
+	dev->mtu = new_mtu;
+	return 0;
+}
+
+#ifdef CONFIG_NET_IPGRE_BROADCAST
+/* Nice toy. Unfortunately, useless in real life :-)
+   It allows to construct virtual multiprotocol broadcast "LAN"
+   over the Internet, provided multicast routing is tuned.
+
+
+   I have no idea was this bicycle invented before me,
+   so that I had to set ARPHRD_IPGRE to a random value.
+   I have an impression, that Cisco could make something similar,
+   but this feature is apparently missing in IOS<=11.2(8).
+   
+   I set up 10.66.66/24 and fec0:6666:6666::0/96 as virtual networks
+   with broadcast 224.66.66.66. If you have access to mbone, play with me :-)
+
+   ping -t 255 224.66.66.66
+
+   If nobody answers, mbone does not work.
+
+   ip tunnel add Universe mode gre remote 224.66.66.66 local <Your_real_addr> ttl 255
+   ip addr add 10.66.66.<somewhat>/24 dev Universe
+   ifconfig Universe up
+   ifconfig Universe add fe80::<Your_real_addr>/10
+   ifconfig Universe add fec0:6666:6666::<Your_real_addr>/96
+   ftp 10.66.66.66
+   ...
+   ftp fec0:6666:6666::193.233.7.65
+   ...
+
+ */
+
+static int ipgre_header(struct sk_buff *skb, struct device *dev, unsigned short type,
+			void *daddr, void *saddr, unsigned len)
+{
+	struct ip_tunnel *t = (struct ip_tunnel*)dev->priv;
+	struct iphdr *iph = (struct iphdr *)skb_push(skb, t->hlen);
+	u16 *p = (u16*)(iph+1);
+
+	memcpy(iph, &t->parms.iph, sizeof(struct iphdr));
+	p[0]		= t->parms.o_flags;
+	p[1]		= htons(type);
+
+	/*
+	 *	Set the source hardware address. 
+	 */
+	 
+	if (saddr)
+		memcpy(&iph->saddr, saddr, 4);
+
+	if (daddr) {
+		memcpy(&iph->daddr, daddr, 4);
+		return t->hlen;
+	}
+	if (iph->daddr && !MULTICAST(iph->daddr))
+		return t->hlen;
+	
+	return -t->hlen;
+}
+
+static int ipgre_rebuild_header(struct sk_buff *skb)
+{
+	struct device *dev = skb->dev;
+	struct iphdr *iph = (struct iphdr *)skb->data;
+	u16 *p = (u16*)(iph + 1);
+ 	struct neighbour *neigh = NULL;
+
+ 	if (skb->dst)
+ 		neigh = skb->dst->neighbour;
+
+ 	if (neigh)
+ 		return neigh->ops->resolve((void*)&iph->daddr, skb);
+
+	if (p[1] == __constant_htons(ETH_P_IP))
+ 		return arp_find((void*)&iph->daddr, skb);
+
+	if (net_ratelimit())
+		printk(KERN_DEBUG "%s: unable to resolve type %X addresses.\n", 
+		       dev->name, (int)p[1]);
+	return 0;
+}
+
+static int ipgre_open(struct device *dev)
+{
+	struct ip_tunnel *t = (struct ip_tunnel*)dev->priv;
+
+	MOD_INC_USE_COUNT;
+	if (MULTICAST(t->parms.iph.daddr)) {
+		struct rtable *rt;
+		if (ip_route_output(&rt, t->parms.iph.daddr,
+				    t->parms.iph.saddr, RT_TOS(t->parms.iph.tos), 
+				    t->parms.link)) {
+			MOD_DEC_USE_COUNT;
+			return -EADDRNOTAVAIL;
+		}
+		dev = rt->u.dst.dev;
+		ip_rt_put(rt);
+		if (dev->ip_ptr == NULL) {
+			MOD_DEC_USE_COUNT;
+			return -EADDRNOTAVAIL;
+		}
+		t->mlink = dev->ifindex;
+		ip_mc_inc_group(dev->ip_ptr, t->parms.iph.daddr);
+	}
+	return 0;
+}
+
+static int ipgre_close(struct device *dev)
+{
+	struct ip_tunnel *t = (struct ip_tunnel*)dev->priv;
+	if (MULTICAST(t->parms.iph.daddr) && t->mlink) {
+		dev = dev_get_by_index(t->mlink);
+		if (dev && dev->ip_ptr)
+			ip_mc_dec_group(dev->ip_ptr, t->parms.iph.daddr);
+	}
+	MOD_DEC_USE_COUNT;
+	return 0;
+}
+
+#endif
+
+static void ipgre_tunnel_init_gen(struct device *dev)
+{
+	struct ip_tunnel *t = (struct ip_tunnel*)dev->priv;
+
+	dev->destructor		= ipgre_tunnel_destroy;
+	dev->hard_start_xmit	= ipgre_tunnel_xmit;
+	dev->get_stats		= ipgre_tunnel_get_stats;
+	dev->do_ioctl		= ipgre_tunnel_ioctl;
+	dev->change_mtu		= ipgre_tunnel_change_mtu;
+
+	dev_init_buffers(dev);
+
+	dev->type		= ARPHRD_IPGRE;
+	dev->hard_header_len 	= LL_MAX_HEADER + sizeof(struct iphdr) + 4;
+	dev->mtu		= 1500 - sizeof(struct iphdr) - 4;
+	dev->flags		= IFF_NOARP;
+	dev->iflink		= 0;
+	dev->addr_len		= 4;
+	memcpy(dev->dev_addr, &t->parms.iph.saddr, 4);
+	memcpy(dev->broadcast, &t->parms.iph.daddr, 4);
+}
+
+static int ipgre_tunnel_init(struct device *dev)
+{
+	struct device *tdev = NULL;
+	struct ip_tunnel *tunnel;
+	struct iphdr *iph;
+	int hlen = LL_MAX_HEADER;
+	int mtu = 1500;
+	int addend = sizeof(struct iphdr) + 4;
+
+	tunnel = (struct ip_tunnel*)dev->priv;
+	iph = &tunnel->parms.iph;
+
+	ipgre_tunnel_init_gen(dev);
+
+	/* Guess output device to choose reasonable mtu and hard_header_len */
+
+	if (iph->daddr) {
+		struct rtable *rt;
+		if (!ip_route_output(&rt, iph->daddr, iph->saddr, RT_TOS(iph->tos), tunnel->parms.link)) {
+			tdev = rt->u.dst.dev;
+			ip_rt_put(rt);
+		}
+
+		dev->flags |= IFF_POINTOPOINT;
+
+#ifdef CONFIG_NET_IPGRE_BROADCAST
+		if (MULTICAST(iph->daddr)) {
+			if (!iph->saddr)
+				return -EINVAL;
+			dev->flags = IFF_BROADCAST;
+			dev->hard_header = ipgre_header;
+			dev->rebuild_header = ipgre_rebuild_header;
+			dev->open = ipgre_open;
+			dev->stop = ipgre_close;
+		}
+#endif
+	}
+
+	if (!tdev && tunnel->parms.link)
+		tdev = dev_get_by_index(tunnel->parms.link);
+
+	if (tdev) {
+		hlen = tdev->hard_header_len;
+		mtu = tdev->mtu;
+	}
+	dev->iflink = tunnel->parms.link;
+
+	/* Precalculate GRE options length */
+	if (tunnel->parms.o_flags&(GRE_CSUM|GRE_KEY|GRE_SEQ)) {
+		if (tunnel->parms.o_flags&GRE_CSUM)
+			addend += 4;
+		if (tunnel->parms.o_flags&GRE_KEY)
+			addend += 4;
+		if (tunnel->parms.o_flags&GRE_SEQ)
+			addend += 4;
+	}
+	dev->hard_header_len = hlen + addend;
+	dev->mtu = mtu - addend;
+	tunnel->hlen = addend;
+	return 0;
+}
+
+#ifdef MODULE
+static int ipgre_fb_tunnel_open(struct device *dev)
+{
+	MOD_INC_USE_COUNT;
+	return 0;
+}
+
+static int ipgre_fb_tunnel_close(struct device *dev)
+{
+	MOD_DEC_USE_COUNT;
+	return 0;
+}
+#endif
+
+__initfunc(int ipgre_fb_tunnel_init(struct device *dev))
+{
+	struct ip_tunnel *tunnel = (struct ip_tunnel*)dev->priv;
+	struct iphdr *iph;
+
+	ipgre_tunnel_init_gen(dev);
+#ifdef MODULE
+	dev->open		= ipgre_fb_tunnel_open;
+	dev->stop		= ipgre_fb_tunnel_close;
+#endif
+
+	iph = &ipgre_fb_tunnel.parms.iph;
+	iph->version		= 4;
+	iph->protocol		= IPPROTO_GRE;
+	iph->ihl		= 5;
+	tunnel->hlen		= sizeof(struct iphdr) + 4;
+
+	tunnels_wc[0]		= &ipgre_fb_tunnel;
+	return 0;
+}
+
+
+static struct inet_protocol ipgre_protocol = {
+  ipgre_rcv,             /* GRE handler          */
+  ipgre_err,             /* TUNNEL error control */
+  0,                    /* next                 */
+  IPPROTO_GRE,          /* protocol ID          */
+  0,                    /* copy                 */
+  NULL,                 /* data                 */
+  "GRE"                 /* name                 */
+};
+
+
+/*
+ *	And now the modules code and kernel interface.
+ */
+
+#ifdef MODULE
+int init_module(void) 
+#else
+__initfunc(int ipgre_init(void))
+#endif
+{
+	printk(KERN_INFO "GRE over IPv4 tunneling driver\n");
+
+	ipgre_fb_tunnel_dev.priv = (void*)&ipgre_fb_tunnel;
+	ipgre_fb_tunnel_dev.name = ipgre_fb_tunnel.parms.name;
+#ifdef MODULE
+	register_netdev(&ipgre_fb_tunnel_dev);
+#else
+	register_netdevice(&ipgre_fb_tunnel_dev);
+#endif
+
+	inet_add_protocol(&ipgre_protocol);
+	return 0;
+}
+
+#ifdef MODULE
+
+void cleanup_module(void)
+{
+	if ( inet_del_protocol(&ipgre_protocol) < 0 )
+		printk(KERN_INFO "ipgre close: can't remove protocol\n");
+
+	unregister_netdev(&ipgre_fb_tunnel_dev);
+}
+
+#endif
diff --git a/net/ipv4/ipconfig.c b/net/ipv4/ipconfig.c
new file mode 100644
index 000000000..30df2360d
--- /dev/null
+++ b/net/ipv4/ipconfig.c
@@ -0,0 +1,1160 @@
+/*
+ *  $Id: ipconfig.c,v 1.5 1997/10/27 16:08:02 mj Exp $
+ *
+ *  Automatic Configuration of IP -- use BOOTP or RARP or user-supplied
+ *  information to configure own IP address and routes.
+ *
+ *  Copyright (C) 1996, 1997 Martin Mares <mj@atrey.karlin.mff.cuni.cz>
+ *
+ *  Derived from network configuration code in fs/nfs/nfsroot.c,
+ *  originally Copyright (C) 1995, 1996 Gero Kuhlmann and me.
+ */
+
+#include <linux/config.h>
+#include <linux/types.h>
+#include <linux/string.h>
+#include <linux/kernel.h>
+#include <linux/sched.h>
+#include <linux/random.h>
+#include <linux/init.h>
+#include <linux/utsname.h>
+#include <linux/in.h>
+#include <linux/if.h>
+#include <linux/inet.h>
+#include <linux/net.h>
+#include <linux/netdevice.h>
+#include <linux/if_arp.h>
+#include <linux/skbuff.h>
+#include <linux/ip.h>
+#include <linux/socket.h>
+#include <linux/inetdevice.h>
+#include <linux/route.h>
+#include <net/route.h>
+#include <net/sock.h>
+#include <net/arp.h>
+#include <net/ip_fib.h>
+#include <net/ipconfig.h>
+
+#include <asm/segment.h>
+#include <asm/uaccess.h>
+
+/* Define this to allow debugging output */
+#undef IPCONFIG_DEBUG
+
+#ifdef IPCONFIG_DEBUG
+#define DBG(x) printk x
+#else
+#define DBG(x) do { } while(0)
+#endif
+
+/* Define the timeout for waiting for a RARP/BOOTP reply */
+#define CONF_BASE_TIMEOUT	(HZ*5)	/* Initial timeout: 5 seconds */
+#define CONF_RETRIES	 	10	/* 10 retries */
+#define CONF_TIMEOUT_RANDOM	(HZ)	/* Maximum amount of randomization */
+#define CONF_TIMEOUT_MULT	*5/4	/* Rate of timeout growth */
+#define CONF_TIMEOUT_MAX	(HZ*30)	/* Maximum allowed timeout */
+
+/* IP configuration */
+static char user_dev_name[IFNAMSIZ] __initdata = { 0, };/* Name of user-selected boot device */
+u32 ic_myaddr __initdata = INADDR_NONE;		/* My IP address */
+u32 ic_servaddr __initdata = INADDR_NONE;	/* Server IP address */
+u32 ic_gateway __initdata = INADDR_NONE;	/* Gateway IP address */
+u32 ic_netmask __initdata = INADDR_NONE;	/* Netmask for local subnet */
+int ic_bootp_flag __initdata = 1;		/* Use BOOTP */
+int ic_rarp_flag __initdata = 1;		/* Use RARP */
+int ic_enable __initdata = 1;			/* Automatic IP configuration enabled */
+int ic_host_name_set __initdata = 0;		/* Host name configured manually */
+int ic_set_manually __initdata = 0;		/* IPconfig parameters set manually */
+
+u32 root_server_addr __initdata = INADDR_NONE;		/* Address of boot server */
+u8 root_server_path[256] __initdata = { 0, };		/* Path to mount as root */
+
+#if defined(CONFIG_IP_PNP_BOOTP) || defined(CONFIG_IP_PNP_RARP)
+
+#define CONFIG_IP_PNP_DYNAMIC
+
+static int ic_got_reply __initdata = 0;
+
+#define IC_GOT_BOOTP 1
+#define IC_GOT_RARP 2
+
+#endif
+
+/*
+ *	Network devices
+ */
+
+struct ic_device {
+	struct ic_device *next;
+	struct device *dev;
+	unsigned short flags;
+};
+
+static struct ic_device *ic_first_dev __initdata = NULL;/* List of open device */
+static struct device *ic_dev __initdata = NULL;		/* Selected device */
+static int bootp_dev_count __initdata = 0;		/* BOOTP capable devices */
+static int rarp_dev_count __initdata = 0;		/* RARP capable devices */
+
+__initfunc(int ic_open_devs(void))
+{
+	struct ic_device *d, **last;
+	struct device *dev;
+	unsigned short oflags;
+
+	last = &ic_first_dev;
+	for (dev = dev_base; dev; dev = dev->next)
+		if (dev->type < ARPHRD_SLIP &&
+		    !(dev->flags & (IFF_LOOPBACK | IFF_POINTOPOINT)) &&
+		    strncmp(dev->name, "dummy", 5) &&
+		    (!user_dev_name[0] || !strcmp(dev->name, user_dev_name))) {
+			oflags = dev->flags;
+			if (dev_change_flags(dev, oflags | IFF_UP) < 0) {
+				printk(KERN_ERR "IP-Config: Failed to open %s\n", dev->name);
+				continue;
+			}
+			if (!(d = kmalloc(sizeof(struct ic_device), GFP_KERNEL)))
+				return -1;
+			d->dev = dev;
+			*last = d;
+			last = &d->next;
+			d->flags = oflags;
+			bootp_dev_count++;
+			if (!(dev->flags & IFF_NOARP))
+				rarp_dev_count++;
+			DBG(("IP-Config: Opened %s\n", dev->name));
+		}
+	*last = NULL;
+
+	if (!bootp_dev_count) {
+		if (user_dev_name[0])
+			printk(KERN_ERR "IP-Config: Device `%s' not found.\n", user_dev_name);
+		else
+			printk(KERN_ERR "IP-Config: No network devices available.\n");
+		return -1;
+	}
+	return 0;
+}
+
+__initfunc(void ic_close_devs(void))
+{
+	struct ic_device *d, *next;
+	struct device *dev;
+
+	next = ic_first_dev;
+	while ((d = next)) {
+		next = d->next;
+		dev = d->dev;
+		if (dev != ic_dev) {
+			DBG(("IP-Config: Downing %s\n", dev->name));
+			dev_change_flags(dev, d->flags);
+		}
+		kfree_s(d, sizeof(struct ic_device));
+	}
+}
+
+/*
+ *	Interface to various network functions.
+ */
+
+static inline void
+set_sockaddr(struct sockaddr_in *sin, u32 addr, u16 port)
+{
+	sin->sin_family = AF_INET;
+	sin->sin_addr.s_addr = addr;
+	sin->sin_port = port;
+}
+
+__initfunc(static int ic_dev_ioctl(unsigned int cmd, struct ifreq *arg))
+{
+	int res;
+
+	mm_segment_t oldfs = get_fs();
+	set_fs(get_ds());
+	res = devinet_ioctl(cmd, arg);
+	set_fs(oldfs);
+	return res;
+}
+
+__initfunc(static int ic_route_ioctl(unsigned int cmd, struct rtentry *arg))
+{
+	int res;
+
+	mm_segment_t oldfs = get_fs();
+	set_fs(get_ds());
+	res = ip_rt_ioctl(cmd, arg);
+	set_fs(oldfs);
+	return res;
+}
+
+/*
+ *	Set up interface addresses and routes.
+ */
+
+__initfunc(static int ic_setup_if(void))
+{
+	struct ifreq ir;
+	struct sockaddr_in *sin = (void *) &ir.ifr_ifru.ifru_addr;
+	int err;
+
+	memset(&ir, 0, sizeof(ir));
+	strcpy(ir.ifr_ifrn.ifrn_name, ic_dev->name);
+	set_sockaddr(sin, ic_myaddr, 0);
+	if ((err = ic_dev_ioctl(SIOCSIFADDR, &ir)) < 0) {
+		printk(KERN_ERR "IP-Config: Unable to set interface address (%d).\n", err);
+		return -1;
+	}
+	set_sockaddr(sin, ic_netmask, 0);
+	if ((err = ic_dev_ioctl(SIOCSIFNETMASK, &ir)) < 0) {
+		printk(KERN_ERR "IP-Config: Unable to set interface netmask (%d).\n", err);
+		return -1;
+	}
+	set_sockaddr(sin, ic_myaddr | ~ic_netmask, 0);
+	if ((err = ic_dev_ioctl(SIOCSIFBRDADDR, &ir)) < 0) {
+		printk(KERN_ERR "IP-Config: Unable to set interface broadcast address (%d).\n", err);
+		return -1;
+	}
+	return 0;
+}
+
+__initfunc(int ic_setup_routes(void))
+{
+	/* No need to setup device routes, only the default route... */
+
+	if (ic_gateway != INADDR_NONE) {
+		struct rtentry rm;
+		int err;
+
+		memset(&rm, 0, sizeof(rm));
+		if ((ic_gateway ^ ic_myaddr) & ic_netmask) {
+			printk(KERN_ERR "IP-Config: Gateway not on directly connected network.\n");
+			return -1;
+		}
+		set_sockaddr((struct sockaddr_in *) &rm.rt_dst, 0, 0);
+		set_sockaddr((struct sockaddr_in *) &rm.rt_genmask, 0, 0);
+		set_sockaddr((struct sockaddr_in *) &rm.rt_gateway, ic_gateway, 0);
+		rm.rt_flags = RTF_UP | RTF_GATEWAY;
+		if ((err = ic_route_ioctl(SIOCADDRT, &rm)) < 0) {
+			printk(KERN_ERR "IP-Config: Cannot add default route (%d).\n", err);
+			return -1;
+		}
+	}
+
+	return 0;
+}
+
+/*
+ *	Fill in default values for all missing parameters.
+ */
+
+__initfunc(int ic_defaults(void))
+{
+	if (!ic_host_name_set)
+		strcpy(system_utsname.nodename, in_ntoa(ic_myaddr));
+
+	if (root_server_addr == INADDR_NONE)
+		root_server_addr = ic_servaddr;
+
+	if (ic_netmask == INADDR_NONE) {
+		if (IN_CLASSA(ic_myaddr))
+			ic_netmask = IN_CLASSA_NET;
+		else if (IN_CLASSB(ic_myaddr))
+			ic_netmask = IN_CLASSB_NET;
+		else if (IN_CLASSC(ic_myaddr))
+			ic_netmask = IN_CLASSC_NET;
+		else {
+			printk(KERN_ERR "IP-Config: Unable to guess netmask for address %08x\n", ic_myaddr);
+			return -1;
+		}
+	}
+
+	return 0;
+}
+
+/*
+ *	RARP support.
+ */
+
+#ifdef CONFIG_IP_PNP_RARP
+
+static int ic_rarp_recv(struct sk_buff *skb, struct device *dev,
+			struct packet_type *pt);
+
+static struct packet_type rarp_packet_type __initdata = {
+	0,			/* Should be: __constant_htons(ETH_P_RARP)
+				 * - but this _doesn't_ come out constant! */
+	NULL,			/* Listen to all devices */
+	ic_rarp_recv,
+	NULL,
+	NULL
+};
+
+__initfunc(static void ic_rarp_init(void))
+{
+	rarp_packet_type.type = htons(ETH_P_RARP);
+	dev_add_pack(&rarp_packet_type);
+}
+
+__initfunc(static void ic_rarp_cleanup(void))
+{
+	dev_remove_pack(&rarp_packet_type);
+}
+
+/*
+ *  Process received RARP packet.
+ */
+__initfunc(static int
+ic_rarp_recv(struct sk_buff *skb, struct device *dev, struct packet_type *pt))
+{
+	struct arphdr *rarp = (struct arphdr *)skb->h.raw;
+	unsigned char *rarp_ptr = (unsigned char *) (rarp + 1);
+	unsigned long sip, tip;
+	unsigned char *sha, *tha;		/* s for "source", t for "target" */
+
+	/* If this test doesn't pass, it's not IP, or we should ignore it anyway */
+	if (rarp->ar_hln != dev->addr_len || dev->type != ntohs(rarp->ar_hrd))
+		goto drop;
+
+	/* If it's not a RARP reply, delete it. */
+	if (rarp->ar_op != htons(ARPOP_RREPLY))
+		goto drop;
+
+	/* If it's not ethernet, delete it. */
+	if (rarp->ar_pro != htons(ETH_P_IP))
+		goto drop;
+
+	/* Extract variable-width fields */
+	sha = rarp_ptr;
+	rarp_ptr += dev->addr_len;
+	memcpy(&sip, rarp_ptr, 4);
+	rarp_ptr += 4;
+	tha = rarp_ptr;
+	rarp_ptr += dev->addr_len;
+	memcpy(&tip, rarp_ptr, 4);
+
+	/* Discard packets which are not meant for us. */
+	if (memcmp(tha, dev->dev_addr, dev->addr_len))
+		goto drop;
+
+	/* Discard packets which are not from specified server. */
+	if (ic_servaddr != INADDR_NONE && ic_servaddr != sip)
+		goto drop;
+
+	/* Victory! The packet is what we were looking for! */
+	if (!ic_got_reply) {
+		ic_got_reply = IC_GOT_RARP;
+		ic_dev = dev;
+		if (ic_myaddr == INADDR_NONE)
+			ic_myaddr = tip;
+		ic_servaddr = sip;
+	}
+
+	/* And throw the packet out... */
+drop:
+	kfree_skb(skb, FREE_READ);
+	return 0;
+}
+
+
+/*
+ *  Send RARP request packet over all devices which allow RARP.
+ */
+__initfunc(static void ic_rarp_send(void))
+{
+	struct ic_device *d;
+
+	for (d=ic_first_dev; d; d=d->next) {
+		struct device *dev = d->dev;
+		if (!(dev->flags & IFF_NOARP))
+			arp_send(ARPOP_RREQUEST, ETH_P_RARP, 0, dev, 0, NULL,
+				 dev->dev_addr, dev->dev_addr);
+	}
+}
+
+#endif
+
+/*
+ *	BOOTP support.
+ */
+
+#ifdef CONFIG_IP_PNP_BOOTP
+
+static struct socket *ic_bootp_xmit_sock __initdata = NULL; /* BOOTP send socket */
+static struct socket *ic_bootp_recv_sock __initdata = NULL; /* BOOTP receive socket */
+
+struct bootp_pkt {		/* BOOTP packet format */
+	u8 op;			/* 1=request, 2=reply */
+	u8 htype;		/* HW address type */
+	u8 hlen;		/* HW address length */
+	u8 hops;		/* Used only by gateways */
+	u32 xid;		/* Transaction ID */
+	u16 secs;		/* Seconds since we started */
+	u16 flags;		/* Just what it says */
+	u32 client_ip;		/* Client's IP address if known */
+	u32 your_ip;		/* Assigned IP address */
+	u32 server_ip;		/* Server's IP address */
+	u32 relay_ip;		/* IP address of BOOTP relay */
+	u8 hw_addr[16];		/* Client's HW address */
+	u8 serv_name[64];	/* Server host name */
+	u8 boot_file[128];	/* Name of boot file */
+	u8 vendor_area[128];	/* Area for extensions */
+};
+
+#define BOOTP_REQUEST 1
+#define BOOTP_REPLY 2
+
+static struct bootp_pkt *ic_xmit_bootp __initdata = NULL; /* Packet being transmitted */
+static struct bootp_pkt *ic_recv_bootp __initdata = NULL; /* Packet being received */
+
+/*
+ *  Dirty tricks for BOOTP packet routing.  We replace the standard lookup function
+ *  for the local fib by our version which does fake lookups and returns our private
+ *  fib entries. Ugly, but it seems to be the simplest way to do the job.
+ */
+
+static void *ic_old_local_lookup __initdata = NULL;	/* Old local routing table lookup function */
+static struct fib_info *ic_bootp_tx_fib __initdata = NULL; /* Our fake fib entries */
+static struct fib_info *ic_bootp_rx_fib __initdata = NULL;
+
+__initfunc(static int ic_bootp_route_lookup(struct fib_table *tb, const struct rt_key *key,
+	struct fib_result *res))
+{
+	static u32 ic_brl_zero = 0;
+
+	DBG(("BOOTP: Route lookup: %d:%08x -> %d:%08x: ", key->iif, key->src, key->oif, key->dst));
+	res->scope = RT_SCOPE_UNIVERSE;
+	res->prefix = &ic_brl_zero;
+	res->prefixlen = 0;
+	res->nh_sel = 0;
+	if (key->src == 0 && key->dst == 0xffffffff && key->iif == loopback_dev.ifindex) { /* Packet output */
+		DBG(("Output\n"));
+		res->type = RTN_UNICAST;
+		res->fi = ic_bootp_tx_fib;
+	} else if (key->iif && key->iif != loopback_dev.ifindex && key->oif == 0) {	/* Packet input */
+		DBG(("Input\n"));
+		res->type = RTN_LOCAL;
+		res->fi = ic_bootp_rx_fib;
+	} else if (!key->iif && !key->oif && !key->src) {	/* Address check by inet_addr_type() */
+		DBG(("Check\n"));
+		res->type = RTN_UNICAST;
+		res->fi = ic_bootp_tx_fib;
+	} else {
+		DBG(("Drop\n"));
+		return -EINVAL;
+	}
+	return 0;
+}
+
+__initfunc(static int ic_set_bootp_route(struct ic_device *d))
+{
+	struct fib_info *f = ic_bootp_tx_fib;
+	struct fib_nh *n = &f->fib_nh[0];
+
+	n->nh_dev = d->dev;
+	n->nh_oif = n->nh_dev->ifindex;
+	rt_cache_flush(0);
+	return 0;
+}
+
+__initfunc(static int ic_bootp_route_init(void))
+{
+	int size = sizeof(struct fib_info) + sizeof(struct fib_nh);
+	struct fib_info *rf, *tf;
+	struct fib_nh *nh;
+
+	if (!(rf = ic_bootp_rx_fib = kmalloc(size, GFP_KERNEL)) ||
+	    !(tf = ic_bootp_tx_fib = kmalloc(size, GFP_KERNEL)))
+		return -1;
+
+	memset(rf, 0, size);
+	rf->fib_nhs = 1;
+	nh = &rf->fib_nh[0];
+	nh->nh_scope = RT_SCOPE_UNIVERSE;
+
+	memset(tf, 0, size);
+	rf->fib_nhs = 1;
+	nh = &rf->fib_nh[0];
+	nh->nh_dev = ic_first_dev->dev;
+	nh->nh_scope = RT_SCOPE_UNIVERSE;
+	nh->nh_oif = nh->nh_dev->ifindex;
+
+	/* Dirty trick: replace standard routing table lookup by our function */
+	ic_old_local_lookup = local_table->tb_lookup;
+	local_table->tb_lookup = ic_bootp_route_lookup;
+
+	return 0;
+}
+
+__initfunc(static void ic_bootp_route_cleanup(void))
+{
+	if (ic_old_local_lookup)
+		local_table->tb_lookup = ic_old_local_lookup;
+	if (ic_bootp_rx_fib)
+		kfree_s(ic_bootp_rx_fib, sizeof(struct fib_info) + sizeof(struct fib_nh));
+	if (ic_bootp_tx_fib)
+		kfree_s(ic_bootp_tx_fib, sizeof(struct fib_info) + sizeof(struct fib_nh));
+}
+
+
+/*
+ *  Allocation and freeing of BOOTP packet buffers.
+ */
+__initfunc(static int ic_bootp_alloc(void))
+{
+	if (!(ic_xmit_bootp = kmalloc(sizeof(struct bootp_pkt), GFP_KERNEL)) ||
+	    !(ic_recv_bootp = kmalloc(sizeof(struct bootp_pkt), GFP_KERNEL))) {
+		printk(KERN_ERR "BOOTP: Out of memory!\n");
+		return -1;
+	}
+	return 0;
+}
+
+__initfunc(static void ic_bootp_free(void))
+{
+	if (ic_xmit_bootp) {
+		kfree_s(ic_xmit_bootp, sizeof(struct bootp_pkt));
+		ic_xmit_bootp = NULL;
+	}
+	if (ic_recv_bootp) {
+		kfree_s(ic_recv_bootp, sizeof(struct bootp_pkt));
+		ic_recv_bootp = NULL;
+	}
+}
+
+
+/*
+ *  Add / Remove fake interface addresses for BOOTP packet sending.
+ */
+__initfunc(static int ic_bootp_addrs_add(void))
+{
+	struct ic_device *d;
+	int err;
+
+	for(d=ic_first_dev; d; d=d->next)
+		if ((err = inet_add_bootp_addr(d->dev)) < 0) {
+			printk(KERN_ERR "BOOTP: Unable to set interface address\n");
+			return -1;
+		}
+	return 0;
+}
+
+__initfunc(static void ic_bootp_addrs_del(void))
+{
+	struct ic_device *d;
+
+	for(d=ic_first_dev; d; d=d->next)
+		inet_del_bootp_addr(d->dev);
+}
+
+/*
+ *  UDP socket operations.
+ */
+__initfunc(static int ic_udp_open(struct socket **sock))
+{
+	int err;
+
+	if ((err = sock_create(AF_INET, SOCK_DGRAM, IPPROTO_UDP, sock)) < 0)
+		printk(KERN_ERR "BOOTP: Cannot open UDP socket!\n");
+	return err;
+}
+
+static inline void ic_udp_close(struct socket *sock)
+{
+	if (sock)
+		sock_release(sock);
+}
+
+__initfunc(static int ic_udp_connect(struct socket *sock, u32 addr, u16 port))
+{
+	struct sockaddr_in sa;
+	int err;
+
+	set_sockaddr(&sa, htonl(addr), htons(port));
+	err = sock->ops->connect(sock, (struct sockaddr *) &sa, sizeof(sa), 0);
+	if (err < 0) {
+		printk(KERN_ERR "BOOTP: connect() failed (%d)\n", err);
+		return -1;
+	}
+	return 0;
+}
+
+__initfunc(static int ic_udp_bind(struct socket *sock, u32 addr, u16 port))
+{
+	struct sockaddr_in sa;
+	int err;
+
+	set_sockaddr(&sa, htonl(addr), htons(port));
+	err = sock->ops->bind(sock, (struct sockaddr *) &sa, sizeof(sa));
+	if (err < 0) {
+		printk(KERN_ERR "BOOTP: bind() failed (%d)\n", err);
+		return -1;
+	}
+	return 0;
+}
+
+__initfunc(static int ic_udp_send(struct socket *sock, void *buf, int size))
+{
+	mm_segment_t oldfs;
+	int result;
+	struct msghdr msg;
+	struct iovec iov;
+
+	oldfs = get_fs();
+	set_fs(get_ds());
+	iov.iov_base = buf;
+	iov.iov_len = size;
+	memset(&msg, 0, sizeof(msg));
+	msg.msg_iov = &iov;
+	msg.msg_iovlen = 1;
+	result = sock_sendmsg(sock, &msg, size);
+	set_fs(oldfs);
+
+	return (result != size);
+}
+
+__initfunc(static int ic_udp_recv(struct socket *sock, void *buf, int size))
+{
+	mm_segment_t oldfs;
+	int result;
+	struct msghdr msg;
+	struct iovec iov;
+
+	oldfs = get_fs();
+	set_fs(get_ds());
+	iov.iov_base = buf;
+	iov.iov_len = size;
+	memset(&msg, 0, sizeof(msg));
+	msg.msg_flags = MSG_DONTWAIT;
+	msg.msg_iov = &iov;
+	msg.msg_iovlen = 1;
+	result = sock_recvmsg(sock, &msg, size, MSG_DONTWAIT);
+	set_fs(oldfs);
+	return result;
+}
+
+
+/*
+ *  Initialize BOOTP extension fields in the request.
+ */
+__initfunc(static void ic_bootp_init_ext(u8 *e))
+{
+	*e++ = 99;		/* RFC1048 Magic Cookie */
+	*e++ = 130;
+	*e++ = 83;
+	*e++ = 99;
+	*e++ = 1;		/* Subnet mask request */
+	*e++ = 4;
+	e += 4;
+	*e++ = 3;		/* Default gateway request */
+	*e++ = 4;
+	e += 4;
+	*e++ = 12;		/* Host name request */
+	*e++ = 32;
+	e += 32;
+	*e++ = 40;		/* NIS Domain name request */
+	*e++ = 32;
+	e += 32;
+	*e++ = 17;		/* Boot path */
+	*e++ = 32;
+	e += 32;
+	*e = 255;		/* End of the list */
+}
+
+
+/*
+ *  Initialize the BOOTP mechanism.
+ */
+__initfunc(static int ic_bootp_init(void))
+{
+	/* Allocate memory for BOOTP packets */
+	if (ic_bootp_alloc() < 0)
+		return -1;
+
+	/* Add fake zero addresses to all interfaces */
+	if (ic_bootp_addrs_add() < 0)
+		return -1;
+
+	/* Initialize BOOTP routing */
+	if (ic_bootp_route_init() < 0)
+		return -1;
+
+	/* Initialize common portion of BOOTP request */
+	memset(ic_xmit_bootp, 0, sizeof(struct bootp_pkt));
+	ic_xmit_bootp->op = BOOTP_REQUEST;
+	get_random_bytes(&ic_xmit_bootp->xid, sizeof(ic_xmit_bootp->xid));
+	ic_bootp_init_ext(ic_xmit_bootp->vendor_area);
+
+	DBG(("BOOTP: XID=%08x\n", ic_xmit_bootp->xid));
+
+	/* Open the sockets */
+	if (ic_udp_open(&ic_bootp_xmit_sock) ||
+	    ic_udp_open(&ic_bootp_recv_sock))
+		return -1;
+
+	/* Bind/connect the sockets */
+	ic_bootp_xmit_sock->sk->broadcast = 1;
+	ic_bootp_xmit_sock->sk->reuse = 1;
+	ic_bootp_recv_sock->sk->reuse = 1;
+	ic_set_bootp_route(ic_first_dev);
+	if (ic_udp_bind(ic_bootp_recv_sock, INADDR_ANY, 68) ||
+	    ic_udp_bind(ic_bootp_xmit_sock, INADDR_ANY, 68) ||
+	    ic_udp_connect(ic_bootp_xmit_sock, INADDR_BROADCAST, 67))
+		return -1;
+
+	return 0;
+}
+
+
+/*
+ *  BOOTP cleanup.
+ */
+__initfunc(static void ic_bootp_cleanup(void))
+{
+	ic_udp_close(ic_bootp_xmit_sock);
+	ic_udp_close(ic_bootp_recv_sock);
+	ic_bootp_addrs_del();
+	ic_bootp_free();
+	ic_bootp_route_cleanup();
+}
+
+
+/*
+ *  Send BOOTP request to single interface.
+ */
+__initfunc(static int ic_bootp_send_if(struct ic_device *d, u32 jiffies))
+{
+	struct device *dev = d->dev;
+	struct bootp_pkt *b = ic_xmit_bootp;
+
+	b->htype = dev->type;
+	b->hlen = dev->addr_len;
+	memset(b->hw_addr, 0, sizeof(b->hw_addr));
+	memcpy(b->hw_addr, dev->dev_addr, dev->addr_len);
+	b->secs = htons(jiffies / HZ);
+	ic_set_bootp_route(d);
+	return ic_udp_send(ic_bootp_xmit_sock, b, sizeof(struct bootp_pkt));
+}
+
+
+/*
+ *  Send BOOTP requests to all interfaces.
+ */
+__initfunc(static int ic_bootp_send(u32 jiffies))
+{
+	struct ic_device *d;
+
+	for(d=ic_first_dev; d; d=d->next)
+		if (ic_bootp_send_if(d, jiffies) < 0)
+			return -1;
+	return 0;
+}
+
+
+/*
+ *  Copy BOOTP-supplied string if not already set.
+ */
+__initfunc(static int ic_bootp_string(char *dest, char *src, int len, int max))
+{
+	if (!len)
+		return 0;
+	if (len > max-1)
+		len = max-1;
+	strncpy(dest, src, len);
+	dest[len] = '\0';
+	return 1;
+}
+
+
+/*
+ *  Process BOOTP extension.
+ */
+__initfunc(static void ic_do_bootp_ext(u8 *ext))
+{
+#ifdef IPCONFIG_DEBUG
+	u8 *c;
+
+	printk("BOOTP: Got extension %02x",*ext);
+	for(c=ext+2; c<ext+2+ext[1]; c++)
+		printk(" %02x", *c);
+	printk("\n");
+#endif
+
+	switch (*ext++) {
+		case 1:		/* Subnet mask */
+			if (ic_netmask == INADDR_NONE)
+				memcpy(&ic_netmask, ext+1, 4);
+			break;
+		case 3:		/* Default gateway */
+			if (ic_gateway == INADDR_NONE)
+				memcpy(&ic_gateway, ext+1, 4);
+			break;
+		case 12:	/* Host name */
+			ic_bootp_string(system_utsname.nodename, ext+1, *ext, __NEW_UTS_LEN);
+			ic_host_name_set = 1;
+			break;
+		case 40:	/* NIS Domain name */
+			ic_bootp_string(system_utsname.domainname, ext+1, *ext, __NEW_UTS_LEN);
+			break;
+		case 17:	/* Root path */
+			if (!root_server_path[0])
+				ic_bootp_string(root_server_path, ext+1, *ext, sizeof(root_server_path));
+			break;
+	}
+}
+
+
+/*
+ *  Receive BOOTP request.
+ */
+__initfunc(static void ic_bootp_recv(void))
+{
+	int len;
+	u8 *ext, *end, *opt;
+	struct ic_device *d;
+	struct bootp_pkt *b = ic_recv_bootp;
+
+	if ((len = ic_udp_recv(ic_bootp_recv_sock, b, sizeof(struct bootp_pkt))) < 0)
+		return;
+
+	/* Check consistency of incoming packet */
+	if (len < 300 ||			/* See RFC 1542:2.1 */
+	    b->op != BOOTP_REPLY ||
+	    b->xid != ic_xmit_bootp->xid) {
+		printk("?");
+		return;
+		}
+
+	/* Find interface this arrived from */
+	for(d=ic_first_dev; d; d=d->next) {
+		struct device *dev = d->dev;
+		if (b->htype == dev->type ||
+		    b->hlen == dev->addr_len ||
+		    !memcmp(b->hw_addr, dev->dev_addr, dev->addr_len))
+			break;
+	}
+	if (!d) {	/* Unknown device */
+		printk("!");
+		return;
+	}
+
+	/* Record BOOTP packet arrival */
+	cli();
+	if (ic_got_reply) {
+		sti();
+		return;
+	}
+	ic_got_reply = IC_GOT_BOOTP;
+	sti();
+	ic_dev = d->dev;
+
+	/* Extract basic fields */
+	ic_myaddr = b->your_ip;
+	ic_servaddr = b->server_ip;
+
+	/* Parse extensions */
+	if (b->vendor_area[0] == 99 &&	/* Check magic cookie */
+	    b->vendor_area[1] == 130 &&
+	    b->vendor_area[2] == 83 &&
+	    b->vendor_area[3] == 99) {
+		ext = &b->vendor_area[4];
+		end = (u8 *) b + len;
+		while (ext < end && *ext != 0xff) {
+			if (*ext == 0)		/* Padding */
+				ext++;
+			else {
+				opt = ext;
+				ext += ext[1] + 2;
+				if (ext <= end)
+					ic_do_bootp_ext(opt);
+			}
+		}
+	}
+}
+
+#endif
+
+
+/*
+ *	Dynamic IP configuration -- BOOTP and RARP.
+ */
+
+#ifdef CONFIG_IP_PNP_DYNAMIC
+
+__initfunc(int ic_dynamic(void))
+{
+	int retries;
+	unsigned long timeout, jiff;
+	unsigned long start_jiffies;
+
+	/*
+	 * If neither BOOTP nor RARP was selected, return with an error. This
+	 * routine gets only called when some pieces of information are mis-
+	 * sing, and without BOOTP and RARP we are not able to get that in-
+	 * formation.
+	 */
+	if (!ic_bootp_flag && !ic_rarp_flag) {
+		printk(KERN_ERR "IP-Config: Incomplete network configuration information.\n");
+		return -1;
+	}
+
+#ifdef CONFIG_IP_PNP_BOOTP
+	if (ic_bootp_flag && !bootp_dev_count) {
+		printk(KERN_ERR "BOOTP: No suitable device found.\n");
+		ic_bootp_flag = 0;
+	}
+#else
+	ic_bootp_flag = 0;
+#endif
+
+#ifdef CONFIG_IP_PNP_RARP
+	if (ic_rarp_flag && !rarp_dev_count) {
+		printk(KERN_ERR "RARP: No suitable device found.\n");
+		ic_rarp_flag = 0;
+	}
+#else
+	ic_rarp_flag = 0;
+#endif
+
+	if (!ic_bootp_flag && !ic_rarp_flag)
+		/* Error message already printed */
+		return -1;
+
+	/*
+	 * Setup RARP and BOOTP protocols
+	 */
+#ifdef CONFIG_IP_PNP_RARP
+	if (ic_rarp_flag)
+		ic_rarp_init();
+#endif
+#ifdef CONFIG_IP_PNP_BOOTP
+	if (ic_bootp_flag && ic_bootp_init() < 0) {
+		ic_bootp_cleanup();
+		return -1;
+	}
+#endif
+
+	/*
+	 * Send requests and wait, until we get an answer. This loop
+	 * seems to be a terrible waste of CPU time, but actually there is
+	 * only one process running at all, so we don't need to use any
+	 * scheduler functions.
+	 * [Actually we could now, but the nothing else running note still 
+	 *  applies.. - AC]
+	 */
+	printk(KERN_NOTICE "Sending %s%s%s requests...",
+		ic_bootp_flag ? "BOOTP" : "",
+		ic_bootp_flag && ic_rarp_flag ? " and " : "",
+		ic_rarp_flag ? "RARP" : "");
+	start_jiffies = jiffies;
+	retries = CONF_RETRIES;
+	get_random_bytes(&timeout, sizeof(timeout));
+	timeout = CONF_BASE_TIMEOUT + (timeout % (unsigned) CONF_TIMEOUT_RANDOM);
+	for(;;) {
+#ifdef CONFIG_IP_PNP_BOOTP
+		if (ic_bootp_flag && ic_bootp_send(jiffies - start_jiffies) < 0) {
+			printk(" BOOTP failed!\n");
+			ic_bootp_cleanup();
+			ic_bootp_flag = 0;
+			if (!ic_rarp_flag)
+				break;
+		}
+#endif
+#ifdef CONFIG_IP_PNP_RARP
+		if (ic_rarp_flag)
+			ic_rarp_send();
+#endif
+		printk(".");
+		jiff = jiffies + timeout;
+		while (jiffies < jiff && !ic_got_reply)
+#ifdef CONFIG_IP_PNP_BOOTP
+			if (ic_bootp_flag)
+				ic_bootp_recv();
+#else
+			;
+#endif
+		if (ic_got_reply) {
+			printk(" OK\n");
+			break;
+		}
+		if (! --retries) {
+			printk(" timed out!\n");
+			break;
+		}
+		timeout = timeout CONF_TIMEOUT_MULT;
+		if (timeout > CONF_TIMEOUT_MAX)
+			timeout = CONF_TIMEOUT_MAX;
+	}
+
+#ifdef CONFIG_IP_PNP_RARP
+	if (ic_rarp_flag)
+		ic_rarp_cleanup();
+#endif
+#ifdef CONFIG_IP_PNP_BOOTP
+	if (ic_bootp_flag)
+		ic_bootp_cleanup();
+#endif
+
+	if (!ic_got_reply)
+		return -1;
+
+	printk("IP-Config: Got %s answer from %s, ",
+		(ic_got_reply == IC_GOT_BOOTP) ? "BOOTP" : "RARP",
+		in_ntoa(ic_servaddr));
+	printk("my address is %s\n", in_ntoa(ic_myaddr));
+
+	return 0;
+}
+
+#endif
+
+/*
+ *	IP Autoconfig dispatcher.
+ */
+
+__initfunc(int ip_auto_config(void))
+{
+	if (!ic_enable)
+		return 0;
+
+	DBG(("IP-Config: Entered.\n"));
+
+	/* Setup all network devices */
+	if (ic_open_devs() < 0)
+		return -1;
+
+	/*
+	 * If the config information is insufficient (e.g., our IP address or
+	 * IP address of the boot server is missing or we have multiple network
+	 * interfaces and no default was set), use BOOTP or RARP to get the
+	 * missing values.
+	 */
+	if (ic_myaddr == INADDR_NONE ||
+#ifdef CONFIG_ROOT_NFS
+	     root_server_addr == INADDR_NONE ||
+#endif
+	    (ic_first_dev && ic_first_dev->next)) {
+#ifdef CONFIG_IP_PNP_DYNAMIC
+		if (ic_dynamic() < 0) {
+			printk(KERN_ERR "IP-Config: Auto-configuration of network failed.\n");
+			ic_close_devs();
+			return -1;
+		}
+#else
+		printk(KERN_ERR "IP-Config: Incomplete network configuration information.\n");
+		ic_close_devs();
+		return -1;
+#endif
+	} else {
+		ic_dev = ic_first_dev->dev;	/* Device selected manually or only one device -> use it */
+	}
+
+	/*
+	 * Use defaults whereever applicable.
+	 */
+	if (ic_defaults() < 0)
+		return -1;
+
+	/*
+	 * Close all network devices except the device we've
+	 * autoconfigured and set up routes.
+	 */
+	ic_close_devs();
+	if (ic_setup_if() < 0 || ic_setup_routes() < 0)
+		return -1;
+
+	DBG(("IP-Config: device=%s, local=%08x, server=%08x, boot=%08x, gw=%08x, mask=%08x\n",
+	    ic_dev->name, ic_myaddr, ic_servaddr, root_server_addr, ic_gateway, ic_netmask));
+	DBG(("IP-Config: host=%s, domain=%s, path=`%s'\n", system_utsname.nodename,
+	    system_utsname.domainname, root_server_path));
+	return 0;
+}
+
+/*
+ *  Decode any IP configuration options in the "ipconfig" kernel command
+ *  line parameter. It consists of option fields separated by colons in
+ *  the following order:
+ *
+ *  <client-ip>:<server-ip>:<gw-ip>:<netmask>:<host name>:<device>:<bootp|rarp>
+ *
+ *  Any of the fields can be empty which means to use a default value:
+ *	<client-ip>	- address given by BOOTP or RARP
+ *	<server-ip>	- address of host returning BOOTP or RARP packet
+ *	<gw-ip>		- none, or the address returned by BOOTP
+ *	<netmask>	- automatically determined from <client-ip>, or the
+ *			  one returned by BOOTP
+ *	<host name>	- <client-ip> in ASCII notation, or the name returned
+ *			  by BOOTP
+ *	<device>	- use all available devices
+ *	<bootp|rarp|both|off> - use both protocols to determine my own address
+ */
+__initfunc(void ip_auto_config_setup(char *addrs, int *ints))
+{
+	char *cp, *ip, *dp;
+	int num = 0;
+
+	ic_set_manually = 1;
+
+	if (!strcmp(addrs, "bootp")) {
+		ic_rarp_flag = 0;
+		return;
+	} else if (!strcmp(addrs, "rarp")) {
+		ic_bootp_flag = 0;
+		return;
+	} else if (!strcmp(addrs, "both")) {
+		return;
+	} else if (!strcmp(addrs, "off")) {
+		ic_enable = 0;
+		return;
+	}
+
+	/* Parse the whole string */
+	ip = addrs;
+	while (ip && *ip) {
+		if ((cp = strchr(ip, ':')))
+			*cp++ = '\0';
+		if (strlen(ip) > 0) {
+			DBG(("IP-Config: Parameter #%d: `%s'\n", num, ip));
+			switch (num) {
+			case 0:
+				if ((ic_myaddr = in_aton(ip)) == INADDR_ANY)
+					ic_myaddr = INADDR_NONE;
+				break;
+			case 1:
+				if ((ic_servaddr = in_aton(ip)) == INADDR_ANY)
+					ic_servaddr = INADDR_NONE;
+				break;
+			case 2:
+				if ((ic_gateway = in_aton(ip)) == INADDR_ANY)
+					ic_gateway = INADDR_NONE;
+				break;
+			case 3:
+				if ((ic_netmask = in_aton(ip)) == INADDR_ANY)
+					ic_netmask = INADDR_NONE;
+				break;
+			case 4:
+				if ((dp = strchr(ip, '.'))) {
+					*dp++ = '\0';
+					strncpy(system_utsname.domainname, dp, __NEW_UTS_LEN);
+					system_utsname.domainname[__NEW_UTS_LEN] = '\0';
+				}
+				strncpy(system_utsname.nodename, ip, __NEW_UTS_LEN);
+				system_utsname.nodename[__NEW_UTS_LEN] = '\0';
+				ic_host_name_set = 1;
+				break;
+			case 5:
+				strncpy(user_dev_name, ip, IFNAMSIZ);
+				user_dev_name[IFNAMSIZ-1] = '\0';
+				break;
+			case 6:
+				if (!strcmp(ip, "rarp"))
+					ic_bootp_flag = 0;
+				else if (!strcmp(ip, "bootp"))
+					ic_rarp_flag = 0;
+				else if (strcmp(ip, "both"))
+					ic_bootp_flag = ic_rarp_flag = 0;
+				break;
+			}
+		}
+		ip = cp;
+		num++;
+	}
+}
diff --git a/net/ipv4/packet.c b/net/ipv4/packet.c
deleted file mode 100644
index e69de29bb..000000000
--- a/net/ipv4/packet.c
+++ /dev/null