ip6_fib.c

ipv6地址转换器

/*
 *	Linux INET6 implementation 
 *	Forwarding Information Database
 *
 *	Authors:
 *	Pedro Roque		<[e-mail]>	
 *
 *	$Id: ip6_fib.c,v 1.17 1999/04/22 10:07:41 davem Exp $
 *
 *	This program is free software; you can redistribute it and/or
 *      modify it under the terms of the GNU General Public License
 *      as published by the Free Software Foundation; either version
 *      2 of the License, or (at your option) any later version.
 */

#include <linux/config.h>
#include <linux/errno.h>
#include <linux/types.h>
#include <linux/net.h>
#include <linux/route.h>
#include <linux/netdevice.h>
#include <linux/in6.h>

#ifdef 	CONFIG_PROC_FS
#include <linux/proc_fs.h>
#endif

#include <net/ipv6.h>
#include <net/ndisc.h>
#include <net/addrconf.h>

#include <net/ip6_fib.h>
#include <net/ip6_route.h>

#define RT6_DEBUG 2
#undef CONFIG_IPV6_SUBTREES

#if RT6_DEBUG >= 1
#define BUG_TRAP(x) ({ if (!(x)) { printk("Assertion (" #x ") failed at " __FILE__ "(%d):" __FUNCTION__ "\n", __LINE__); } })
#else
#define BUG_TRAP(x) do { ; } while (0)
#endif

#if RT6_DEBUG >= 3
#define RT6_TRACE(x...) printk(KERN_DEBUG x)
#else
#define RT6_TRACE(x...) do { ; } while (0)
#endif

struct rt6_statistics	rt6_stats;

enum fib_walk_state_t
{
#ifdef CONFIG_IPV6_SUBTREES
	FWS_S,
#endif
	FWS_L,
	FWS_R,
	FWS_C,
	FWS_U
};

struct fib6_cleaner_t
{
	struct fib6_walker_t w;
	int (*func)(struct rt6_info *, void *arg);
	void *arg;
};

#ifdef CONFIG_IPV6_SUBTREES
#define FWS_INIT FWS_S
#define SUBTREE(fn) ((fn)->subtree)
#else
#define FWS_INIT FWS_L
#define SUBTREE(fn) NULL
#endif

static void fib6_prune_clones(struct fib6_node *fn, struct rt6_info *rt);
static void fib6_repair_tree(struct fib6_node *fn);

/*
 *	A routing update causes an increase of the serial number on the
 *	afected subtree. This allows for cached routes to be asynchronously
 *	tested when modifications are made to the destination cache as a
 *	result of redirects, path MTU changes, etc.
 */

static __u32	rt_sernum	= 0;

static struct timer_list ip6_fib_timer = {
	NULL, NULL,
	0,
	~0UL,
	fib6_run_gc
};

static struct fib6_walker_t fib6_walker_list = {
	&fib6_walker_list, &fib6_walker_list, 
};

#define FOR_WALKERS(w) for ((w)=fib6_walker_list.next; (w) != &fib6_walker_list; (w)=(w)->next)

static __inline__ u32 fib6_new_sernum(void)
{
	u32 n = ++rt_sernum;
	if ((__s32)n <= 0)
		rt_sernum = n = 1;
	return n;
}

/*
 *	Auxiliary address test functions for the radix tree.
 *
 *	These assume a 32bit processor (although it will work on 
 *	64bit processors)
 */

/*
 *	compare "prefix length" bits of an address
 */

static __inline__ int addr_match(void *token1, void *token2, int prefixlen)
{
	__u32 *a1 = token1;
	__u32 *a2 = token2;
	int pdw;
	int pbi;

	pdw = prefixlen >> 5;	  /* num of whole __u32 in prefix */
	pbi = prefixlen &  0x1f;  /* num of bits in incomplete u32 in prefix */

	if (pdw)
		if (memcmp(a1, a2, pdw << 2))
			return 0;

	if (pbi) {
		__u32 mask;

		mask = htonl((0xffffffff) << (32 - pbi));

		if ((a1[pdw] ^ a2[pdw]) & mask)
			return 0;
	}

	return 1;
}

/*
 *	test bit
 */

static __inline__ int addr_bit_set(void *token, int fn_bit)
{
	__u32 *addr = token;

	return htonl(1 << ((~fn_bit)&0x1F)) & addr[fn_bit>>5];
}

/*
 *	find the first different bit between two addresses
 *	length of address must be a multiple of 32bits
 */

static __inline__ int addr_diff(void *token1, void *token2, int addrlen)
{
	__u32 *a1 = token1;
	__u32 *a2 = token2;
	int i;

	addrlen >>= 2;

	for (i = 0; i < addrlen; i++) {
		__u32 xb;

		xb = a1[i] ^ a2[i];

		if (xb) {
			int j = 31;

			xb = ntohl(xb);

			while (test_bit(j, &xb) == 0)
				j--;

			return (i * 32 + 31 - j);
		}
	}

	/*
	 *	we should *never* get to this point since that 
	 *	would mean the addrs are equal
	 *
	 *	However, we do get to it 8) And exacly, when
	 *	addresses are equal 8)
	 *
	 *	ip route add 1111::/128 via ...
	 *	ip route add 1111::/64 via ...
	 *	and we are here.
	 *
	 *	Ideally, this function should stop comparison
	 *	at prefix length. It does not, but it is still OK,
	 *	if returned value is greater than prefix length.
	 *					--ANK (980803)
	 */

	return addrlen<<5;
}

static __inline__ struct fib6_node * node_alloc(void)
{
	struct fib6_node *fn;

	if ((fn = kmalloc(sizeof(struct fib6_node), GFP_ATOMIC)) != NULL) {
		memset(fn, 0, sizeof(struct fib6_node));
		rt6_stats.fib_nodes++;
	}

	return fn;
}

static __inline__ void node_free(struct fib6_node * fn)
{
	rt6_stats.fib_nodes--;
	kfree(fn);
}

static __inline__ void rt6_release(struct rt6_info *rt)
{
	if (atomic_dec_and_test(&rt->rt6i_ref))
		dst_free(&rt->u.dst);
}


/*
 *	Routing Table
 *
 *	return the apropriate node for a routing tree "add" operation
 *	by either creating and inserting or by returning an existing
 *	node.
 */

static struct fib6_node * fib6_add_1(struct fib6_node *root, void *addr,
				     int addrlen, int plen,
				     int offset)
{
	struct fib6_node *fn, *in, *ln;
	struct fib6_node *pn = NULL;
	struct rt6key *key;
	int	bit;
       	int	dir = 0;
	__u32	sernum = fib6_new_sernum();

	RT6_TRACE("fib6_add_1\n");

	/* insert node in tree */

	fn = root;

	if (plen == 0)
		return fn;

	do {
		key = (struct rt6key *)((u8 *)fn->leaf + offset);

		/*
		 *	Prefix match
		 */
		if (plen < fn->fn_bit ||
		    !addr_match(&key->addr, addr, fn->fn_bit))
			goto insert_above;
		
		/*
		 *	Exact match ?
		 */
			 
		if (plen == fn->fn_bit) {
			/* clean up an intermediate node */
			if ((fn->fn_flags & RTN_RTINFO) == 0) {
				rt6_release(fn->leaf);
				fn->leaf = NULL;
			}
			
			fn->fn_sernum = sernum;
				
			return fn;
		}

		/*
		 *	We have more bits to go
		 */
			 
		/* Try to walk down on tree. */
		fn->fn_sernum = sernum;
		dir = addr_bit_set(addr, fn->fn_bit);
		pn = fn;
		fn = dir ? fn->right: fn->left;
	} while (fn);

	/*
	 *	We wlaked to the bottom of tree.
	 *	Create new leaf node without children.
	 */

	ln = node_alloc();

	if (ln == NULL)
		return NULL;
	ln->fn_bit = plen;
			
	ln->parent = pn;
	ln->fn_sernum = sernum;

	if (dir)
		pn->right = ln;
	else
		pn->left  = ln;

	return ln;


insert_above:
	/*
	 * split since we don't have a common prefix anymore or 
	 * we have a less significant route.
	 * we've to insert an intermediate node on the list
	 * this new node will point to the one we need to create
	 * and the current
	 */

	pn = fn->parent;

	/* find 1st bit in difference between the 2 addrs.

	   See comment in addr_diff: bit may be an invalid value,
	   but if it is >= plen, the value is ignored in any case.
	 */
	
	bit = addr_diff(addr, &key->addr, addrlen);

	/* 
	 *		(intermediate)[in]	
	 *	          /	   \
	 *	(new leaf node)[ln] (old node)[fn]
	 */
	if (plen > bit) {
		in = node_alloc();
		ln = node_alloc();
		
		if (in == NULL || ln == NULL) {
			if (in)
				node_free(in);
			if (ln)
				node_free(ln);
			return NULL;
		}

		/* 
		 * new intermediate node. 
		 * RTN_RTINFO will
		 * be off since that an address that chooses one of
		 * the branches would not match less specific routes
		 * in the other branch
		 */

		in->fn_bit = bit;

		in->parent = pn;
		in->leaf = fn->leaf;
		atomic_inc(&in->leaf->rt6i_ref);

		in->fn_sernum = sernum;

		/* update parent pointer */
		if (dir)
			pn->right = in;
		else
			pn->left  = in;

		ln->fn_bit = plen;

		ln->parent = in;
		fn->parent = in;

		ln->fn_sernum = sernum;

		if (addr_bit_set(addr, bit)) {
			in->right = ln;
			in->left  = fn;
		} else {
			in->left  = ln;
			in->right = fn;
		}
	} else { /* plen <= bit */

		/* 
		 *		(new leaf node)[ln]
		 *	          /	   \
		 *	     (old node)[fn] NULL
		 */

		ln = node_alloc();

		if (ln == NULL)
			return NULL;

		ln->fn_bit = plen;

		ln->parent = pn;

		ln->fn_sernum = sernum;
		
		if (dir)
			pn->right = ln;
		else
			pn->left  = ln;

		if (addr_bit_set(&key->addr, plen))
			ln->right = fn;
		else
			ln->left  = fn;

		fn->parent = ln;
	}
	return ln;
}

/*
 *	Insert routing information in a node.
 */

static int fib6_add_rt2node(struct fib6_node *fn, struct rt6_info *rt)
{
	struct rt6_info *iter = NULL;
	struct rt6_info **ins;

	ins = &fn->leaf;

	for (iter = fn->leaf; iter; iter=iter->u.next) {
		/*
		 *	Search for duplicates
		 */

		if (iter->rt6i_metric == rt->rt6i_metric) {
			/*
			 *	Same priority level
			 */

			if ((iter->rt6i_dev == rt->rt6i_dev) &&
			    (iter->rt6i_flowr == rt->rt6i_flowr) &&
			    (ipv6_addr_cmp(&iter->rt6i_gateway,
					   &rt->rt6i_gateway) == 0)) {
				if (!(iter->rt6i_flags&RTF_EXPIRES))
					return -EEXIST;
				iter->rt6i_expires = rt->rt6i_expires;
				if (!(rt->rt6i_flags&RTF_EXPIRES)) {
					iter->rt6i_flags &= ~RTF_EXPIRES;
					iter->rt6i_expires = 0;
				}
				return -EEXIST;
			}
		}

		if (iter->rt6i_metric > rt->rt6i_metric)
			break;

		ins = &iter->u.next;
	}

	/*
	 *	insert node
	 */

	rt->u.next = iter;
	*ins = rt;
	rt->rt6i_node = fn;
	atomic_inc(&rt->rt6i_ref);
#ifdef CONFIG_RTNETLINK
	inet6_rt_notify(RTM_NEWROUTE, rt);
#endif
	rt6_stats.fib_rt_entries++;

	if ((fn->fn_flags & RTN_RTINFO) == 0) {
		rt6_stats.fib_route_nodes++;
		fn->fn_flags |= RTN_RTINFO;
	}

	return 0;
}

static __inline__ void fib6_start_gc(struct rt6_info *rt)
{
	if (ip6_fib_timer.expires == 0 &&
	    (rt->rt6i_flags & (RTF_EXPIRES|RTF_CACHE))) {
		del_timer(&ip6_fib_timer);
		ip6_fib_timer.expires = jiffies + ip6_rt_gc_interval;
		add_timer(&ip6_fib_timer);
	}
}

/*
 *	Add routing information to the routing tree.
 *	<destination addr>/<source addr>
 *	with source addr info in sub-trees
 */

int fib6_add(struct fib6_node *root, struct rt6_info *rt)
{
	struct fib6_node *fn;
	int err = -ENOMEM;

	fn = fib6_add_1(root, &rt->rt6i_dst.addr, sizeof(struct in6_addr),
			rt->rt6i_dst.plen, (u8*) &rt->rt6i_dst - (u8*) rt);

	if (fn == NULL)
		return -ENOMEM;

#ifdef CONFIG_IPV6_SUBTREES
	if (rt->rt6i_src.plen) {
		struct fib6_node *sn;

		if (fn->subtree == NULL) {
			struct fib6_node *sfn;

			/*
			 * Create subtree.
			 *
			 *		fn[main tree]
			 *		|
			 *		sfn[subtree root]
			 *		   \
			 *		    sn[new leaf node]
			 */

			/* Create subtree root node */
			sfn = node_alloc();
			if (sfn == NULL)
				goto st_failure;

			sfn->leaf = &ip6_null_entry;
			atomic_inc(&ip6_null_entry.rt6i_ref);
			sfn->fn_flags = RTN_ROOT;
			sfn->fn_sernum = fib6_new_sernum();

			/* Now add the first leaf node to new subtree */

			sn = fib6_add_1(sfn, &rt->rt6i_src.addr,
					sizeof(struct in6_addr), rt->rt6i_src.plen,
					(u8*) &rt->rt6i_src - (u8*) rt);

			if (sn == NULL) {
				/* If it is failed, discard just allocated
				   root, and then (in st_failure) stale node
				   in main tree.
				 */
				node_free(sfn);
				goto st_failure;
			}

			/* Now link new subtree to main tree */
			sfn->parent = fn;
			fn->subtree = sfn;
			if (fn->leaf == NULL) {
				fn->leaf = rt;
				atomic_inc(&rt->rt6i_ref);
			}
		} else {
			sn = fib6_add_1(fn->subtree, &rt->rt6i_src.addr,
					sizeof(struct in6_addr), rt->rt6i_src.plen,
					(u8*) &rt->rt6i_src - (u8*) rt);

			if (sn == NULL)
				goto st_failure;
		}

		fn = sn;
	}
#endif

	err = fib6_add_rt2node(fn, rt);

	if (err == 0) {
		fib6_start_gc(rt);
		if (!(rt->rt6i_flags&RTF_CACHE))
			fib6_prune_clones(fn, rt);
	}

	if (err)
		dst_free(&rt->u.dst);
	return err;

#ifdef CONFIG_IPV6_SUBTREES
	/* Subtree creation failed, probably main tree node
	   is orphan. If it is, shot it.
	 */
st_failure:
	if (fn && !(fn->fn_flags&RTN_RTINFO|RTN_ROOT))
		fib_repair_tree(fn);
	dst_free(&rt->u.dst);
	return err;
#endif
}

/*