/*
|
/*
|
* Linux INET6 implementation
|
* Linux INET6 implementation
|
* Forwarding Information Database
|
* Forwarding Information Database
|
*
|
*
|
* Authors:
|
* Authors:
|
* Pedro Roque <roque@di.fc.ul.pt>
|
* Pedro Roque <roque@di.fc.ul.pt>
|
*
|
*
|
* $Id: ip6_fib.c,v 1.1.1.1 2004-04-15 01:14:32 phoenix Exp $
|
* $Id: ip6_fib.c,v 1.1.1.1 2004-04-15 01:14:32 phoenix Exp $
|
*
|
*
|
* This program is free software; you can redistribute it and/or
|
* This program is free software; you can redistribute it and/or
|
* modify it under the terms of the GNU General Public License
|
* modify it under the terms of the GNU General Public License
|
* as published by the Free Software Foundation; either version
|
* as published by the Free Software Foundation; either version
|
* 2 of the License, or (at your option) any later version.
|
* 2 of the License, or (at your option) any later version.
|
*/
|
*/
|
|
|
/*
|
/*
|
* Changes:
|
* Changes:
|
* Yuji SEKIYA @USAGI: Support default route on router node;
|
* Yuji SEKIYA @USAGI: Support default route on router node;
|
* remove ip6_null_entry from the top of
|
* remove ip6_null_entry from the top of
|
* routing table.
|
* routing table.
|
*/
|
*/
|
#include <linux/config.h>
|
#include <linux/config.h>
|
#include <linux/errno.h>
|
#include <linux/errno.h>
|
#include <linux/types.h>
|
#include <linux/types.h>
|
#include <linux/net.h>
|
#include <linux/net.h>
|
#include <linux/route.h>
|
#include <linux/route.h>
|
#include <linux/netdevice.h>
|
#include <linux/netdevice.h>
|
#include <linux/in6.h>
|
#include <linux/in6.h>
|
#include <linux/init.h>
|
#include <linux/init.h>
|
|
|
#ifdef CONFIG_PROC_FS
|
#ifdef CONFIG_PROC_FS
|
#include <linux/proc_fs.h>
|
#include <linux/proc_fs.h>
|
#endif
|
#endif
|
|
|
#include <net/ipv6.h>
|
#include <net/ipv6.h>
|
#include <net/ndisc.h>
|
#include <net/ndisc.h>
|
#include <net/addrconf.h>
|
#include <net/addrconf.h>
|
|
|
#include <net/ip6_fib.h>
|
#include <net/ip6_fib.h>
|
#include <net/ip6_route.h>
|
#include <net/ip6_route.h>
|
|
|
#define RT6_DEBUG 2
|
#define RT6_DEBUG 2
|
#undef CONFIG_IPV6_SUBTREES
|
#undef CONFIG_IPV6_SUBTREES
|
|
|
#if RT6_DEBUG >= 3
|
#if RT6_DEBUG >= 3
|
#define RT6_TRACE(x...) printk(KERN_DEBUG x)
|
#define RT6_TRACE(x...) printk(KERN_DEBUG x)
|
#else
|
#else
|
#define RT6_TRACE(x...) do { ; } while (0)
|
#define RT6_TRACE(x...) do { ; } while (0)
|
#endif
|
#endif
|
|
|
struct rt6_statistics rt6_stats;
|
struct rt6_statistics rt6_stats;
|
|
|
static kmem_cache_t * fib6_node_kmem;
|
static kmem_cache_t * fib6_node_kmem;
|
|
|
enum fib_walk_state_t
|
enum fib_walk_state_t
|
{
|
{
|
#ifdef CONFIG_IPV6_SUBTREES
|
#ifdef CONFIG_IPV6_SUBTREES
|
FWS_S,
|
FWS_S,
|
#endif
|
#endif
|
FWS_L,
|
FWS_L,
|
FWS_R,
|
FWS_R,
|
FWS_C,
|
FWS_C,
|
FWS_U
|
FWS_U
|
};
|
};
|
|
|
struct fib6_cleaner_t
|
struct fib6_cleaner_t
|
{
|
{
|
struct fib6_walker_t w;
|
struct fib6_walker_t w;
|
int (*func)(struct rt6_info *, void *arg);
|
int (*func)(struct rt6_info *, void *arg);
|
void *arg;
|
void *arg;
|
};
|
};
|
|
|
rwlock_t fib6_walker_lock = RW_LOCK_UNLOCKED;
|
rwlock_t fib6_walker_lock = RW_LOCK_UNLOCKED;
|
|
|
|
|
#ifdef CONFIG_IPV6_SUBTREES
|
#ifdef CONFIG_IPV6_SUBTREES
|
#define FWS_INIT FWS_S
|
#define FWS_INIT FWS_S
|
#define SUBTREE(fn) ((fn)->subtree)
|
#define SUBTREE(fn) ((fn)->subtree)
|
#else
|
#else
|
#define FWS_INIT FWS_L
|
#define FWS_INIT FWS_L
|
#define SUBTREE(fn) NULL
|
#define SUBTREE(fn) NULL
|
#endif
|
#endif
|
|
|
static void fib6_prune_clones(struct fib6_node *fn, struct rt6_info *rt);
|
static void fib6_prune_clones(struct fib6_node *fn, struct rt6_info *rt);
|
static struct fib6_node * fib6_repair_tree(struct fib6_node *fn);
|
static struct fib6_node * fib6_repair_tree(struct fib6_node *fn);
|
|
|
/*
|
/*
|
* A routing update causes an increase of the serial number on the
|
* A routing update causes an increase of the serial number on the
|
* afected subtree. This allows for cached routes to be asynchronously
|
* afected subtree. This allows for cached routes to be asynchronously
|
* tested when modifications are made to the destination cache as a
|
* tested when modifications are made to the destination cache as a
|
* result of redirects, path MTU changes, etc.
|
* result of redirects, path MTU changes, etc.
|
*/
|
*/
|
|
|
static __u32 rt_sernum = 0;
|
static __u32 rt_sernum = 0;
|
|
|
static struct timer_list ip6_fib_timer = { function: fib6_run_gc };
|
static struct timer_list ip6_fib_timer = { function: fib6_run_gc };
|
|
|
static struct fib6_walker_t fib6_walker_list = {
|
static struct fib6_walker_t fib6_walker_list = {
|
&fib6_walker_list, &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)
|
#define FOR_WALKERS(w) for ((w)=fib6_walker_list.next; (w) != &fib6_walker_list; (w)=(w)->next)
|
|
|
static __inline__ u32 fib6_new_sernum(void)
|
static __inline__ u32 fib6_new_sernum(void)
|
{
|
{
|
u32 n = ++rt_sernum;
|
u32 n = ++rt_sernum;
|
if ((__s32)n <= 0)
|
if ((__s32)n <= 0)
|
rt_sernum = n = 1;
|
rt_sernum = n = 1;
|
return n;
|
return n;
|
}
|
}
|
|
|
/*
|
/*
|
* Auxiliary address test functions for the radix tree.
|
* Auxiliary address test functions for the radix tree.
|
*
|
*
|
* These assume a 32bit processor (although it will work on
|
* These assume a 32bit processor (although it will work on
|
* 64bit processors)
|
* 64bit processors)
|
*/
|
*/
|
|
|
/*
|
/*
|
* compare "prefix length" bits of an address
|
* compare "prefix length" bits of an address
|
*/
|
*/
|
|
|
static __inline__ int addr_match(void *token1, void *token2, int prefixlen)
|
static __inline__ int addr_match(void *token1, void *token2, int prefixlen)
|
{
|
{
|
__u32 *a1 = token1;
|
__u32 *a1 = token1;
|
__u32 *a2 = token2;
|
__u32 *a2 = token2;
|
int pdw;
|
int pdw;
|
int pbi;
|
int pbi;
|
|
|
pdw = prefixlen >> 5; /* num of whole __u32 in prefix */
|
pdw = prefixlen >> 5; /* num of whole __u32 in prefix */
|
pbi = prefixlen & 0x1f; /* num of bits in incomplete u32 in prefix */
|
pbi = prefixlen & 0x1f; /* num of bits in incomplete u32 in prefix */
|
|
|
if (pdw)
|
if (pdw)
|
if (memcmp(a1, a2, pdw << 2))
|
if (memcmp(a1, a2, pdw << 2))
|
return 0;
|
return 0;
|
|
|
if (pbi) {
|
if (pbi) {
|
__u32 mask;
|
__u32 mask;
|
|
|
mask = htonl((0xffffffff) << (32 - pbi));
|
mask = htonl((0xffffffff) << (32 - pbi));
|
|
|
if ((a1[pdw] ^ a2[pdw]) & mask)
|
if ((a1[pdw] ^ a2[pdw]) & mask)
|
return 0;
|
return 0;
|
}
|
}
|
|
|
return 1;
|
return 1;
|
}
|
}
|
|
|
/*
|
/*
|
* test bit
|
* test bit
|
*/
|
*/
|
|
|
static __inline__ int addr_bit_set(void *token, int fn_bit)
|
static __inline__ int addr_bit_set(void *token, int fn_bit)
|
{
|
{
|
__u32 *addr = token;
|
__u32 *addr = token;
|
|
|
return htonl(1 << ((~fn_bit)&0x1F)) & addr[fn_bit>>5];
|
return htonl(1 << ((~fn_bit)&0x1F)) & addr[fn_bit>>5];
|
}
|
}
|
|
|
/*
|
/*
|
* find the first different bit between two addresses
|
* find the first different bit between two addresses
|
* length of address must be a multiple of 32bits
|
* length of address must be a multiple of 32bits
|
*/
|
*/
|
|
|
static __inline__ int addr_diff(void *token1, void *token2, int addrlen)
|
static __inline__ int addr_diff(void *token1, void *token2, int addrlen)
|
{
|
{
|
__u32 *a1 = token1;
|
__u32 *a1 = token1;
|
__u32 *a2 = token2;
|
__u32 *a2 = token2;
|
int i;
|
int i;
|
|
|
addrlen >>= 2;
|
addrlen >>= 2;
|
|
|
for (i = 0; i < addrlen; i++) {
|
for (i = 0; i < addrlen; i++) {
|
__u32 xb;
|
__u32 xb;
|
|
|
xb = a1[i] ^ a2[i];
|
xb = a1[i] ^ a2[i];
|
|
|
if (xb) {
|
if (xb) {
|
int j = 31;
|
int j = 31;
|
|
|
xb = ntohl(xb);
|
xb = ntohl(xb);
|
|
|
while ((xb & (1 << j)) == 0)
|
while ((xb & (1 << j)) == 0)
|
j--;
|
j--;
|
|
|
return (i * 32 + 31 - j);
|
return (i * 32 + 31 - j);
|
}
|
}
|
}
|
}
|
|
|
/*
|
/*
|
* we should *never* get to this point since that
|
* we should *never* get to this point since that
|
* would mean the addrs are equal
|
* would mean the addrs are equal
|
*
|
*
|
* However, we do get to it 8) And exacly, when
|
* However, we do get to it 8) And exacly, when
|
* addresses are equal 8)
|
* addresses are equal 8)
|
*
|
*
|
* ip route add 1111::/128 via ...
|
* ip route add 1111::/128 via ...
|
* ip route add 1111::/64 via ...
|
* ip route add 1111::/64 via ...
|
* and we are here.
|
* and we are here.
|
*
|
*
|
* Ideally, this function should stop comparison
|
* Ideally, this function should stop comparison
|
* at prefix length. It does not, but it is still OK,
|
* at prefix length. It does not, but it is still OK,
|
* if returned value is greater than prefix length.
|
* if returned value is greater than prefix length.
|
* --ANK (980803)
|
* --ANK (980803)
|
*/
|
*/
|
|
|
return addrlen<<5;
|
return addrlen<<5;
|
}
|
}
|
|
|
static __inline__ struct fib6_node * node_alloc(void)
|
static __inline__ struct fib6_node * node_alloc(void)
|
{
|
{
|
struct fib6_node *fn;
|
struct fib6_node *fn;
|
|
|
if ((fn = kmem_cache_alloc(fib6_node_kmem, SLAB_ATOMIC)) != NULL)
|
if ((fn = kmem_cache_alloc(fib6_node_kmem, SLAB_ATOMIC)) != NULL)
|
memset(fn, 0, sizeof(struct fib6_node));
|
memset(fn, 0, sizeof(struct fib6_node));
|
|
|
return fn;
|
return fn;
|
}
|
}
|
|
|
static __inline__ void node_free(struct fib6_node * fn)
|
static __inline__ void node_free(struct fib6_node * fn)
|
{
|
{
|
kmem_cache_free(fib6_node_kmem, fn);
|
kmem_cache_free(fib6_node_kmem, fn);
|
}
|
}
|
|
|
static __inline__ void rt6_release(struct rt6_info *rt)
|
static __inline__ void rt6_release(struct rt6_info *rt)
|
{
|
{
|
if (atomic_dec_and_test(&rt->rt6i_ref))
|
if (atomic_dec_and_test(&rt->rt6i_ref))
|
dst_free(&rt->u.dst);
|
dst_free(&rt->u.dst);
|
}
|
}
|
|
|
|
|
/*
|
/*
|
* Routing Table
|
* Routing Table
|
*
|
*
|
* return the apropriate node for a routing tree "add" operation
|
* return the apropriate node for a routing tree "add" operation
|
* by either creating and inserting or by returning an existing
|
* by either creating and inserting or by returning an existing
|
* node.
|
* node.
|
*/
|
*/
|
|
|
static struct fib6_node * fib6_add_1(struct fib6_node *root, void *addr,
|
static struct fib6_node * fib6_add_1(struct fib6_node *root, void *addr,
|
int addrlen, int plen,
|
int addrlen, int plen,
|
int offset)
|
int offset)
|
{
|
{
|
struct fib6_node *fn, *in, *ln;
|
struct fib6_node *fn, *in, *ln;
|
struct fib6_node *pn = NULL;
|
struct fib6_node *pn = NULL;
|
struct rt6key *key;
|
struct rt6key *key;
|
int bit;
|
int bit;
|
int dir = 0;
|
int dir = 0;
|
__u32 sernum = fib6_new_sernum();
|
__u32 sernum = fib6_new_sernum();
|
|
|
RT6_TRACE("fib6_add_1\n");
|
RT6_TRACE("fib6_add_1\n");
|
|
|
/* insert node in tree */
|
/* insert node in tree */
|
|
|
fn = root;
|
fn = root;
|
|
|
do {
|
do {
|
key = (struct rt6key *)((u8 *)fn->leaf + offset);
|
key = (struct rt6key *)((u8 *)fn->leaf + offset);
|
|
|
/*
|
/*
|
* Prefix match
|
* Prefix match
|
*/
|
*/
|
if (plen < fn->fn_bit ||
|
if (plen < fn->fn_bit ||
|
!addr_match(&key->addr, addr, fn->fn_bit))
|
!addr_match(&key->addr, addr, fn->fn_bit))
|
goto insert_above;
|
goto insert_above;
|
|
|
/*
|
/*
|
* Exact match ?
|
* Exact match ?
|
*/
|
*/
|
|
|
if (plen == fn->fn_bit) {
|
if (plen == fn->fn_bit) {
|
/* clean up an intermediate node */
|
/* clean up an intermediate node */
|
if ((fn->fn_flags & RTN_RTINFO) == 0) {
|
if ((fn->fn_flags & RTN_RTINFO) == 0) {
|
rt6_release(fn->leaf);
|
rt6_release(fn->leaf);
|
fn->leaf = NULL;
|
fn->leaf = NULL;
|
}
|
}
|
|
|
fn->fn_sernum = sernum;
|
fn->fn_sernum = sernum;
|
|
|
return fn;
|
return fn;
|
}
|
}
|
|
|
/*
|
/*
|
* We have more bits to go
|
* We have more bits to go
|
*/
|
*/
|
|
|
/* Try to walk down on tree. */
|
/* Try to walk down on tree. */
|
fn->fn_sernum = sernum;
|
fn->fn_sernum = sernum;
|
dir = addr_bit_set(addr, fn->fn_bit);
|
dir = addr_bit_set(addr, fn->fn_bit);
|
pn = fn;
|
pn = fn;
|
fn = dir ? fn->right: fn->left;
|
fn = dir ? fn->right: fn->left;
|
} while (fn);
|
} while (fn);
|
|
|
/*
|
/*
|
* We walked to the bottom of tree.
|
* We walked to the bottom of tree.
|
* Create new leaf node without children.
|
* Create new leaf node without children.
|
*/
|
*/
|
|
|
ln = node_alloc();
|
ln = node_alloc();
|
|
|
if (ln == NULL)
|
if (ln == NULL)
|
return NULL;
|
return NULL;
|
ln->fn_bit = plen;
|
ln->fn_bit = plen;
|
|
|
ln->parent = pn;
|
ln->parent = pn;
|
ln->fn_sernum = sernum;
|
ln->fn_sernum = sernum;
|
|
|
if (dir)
|
if (dir)
|
pn->right = ln;
|
pn->right = ln;
|
else
|
else
|
pn->left = ln;
|
pn->left = ln;
|
|
|
return ln;
|
return ln;
|
|
|
|
|
insert_above:
|
insert_above:
|
/*
|
/*
|
* split since we don't have a common prefix anymore or
|
* split since we don't have a common prefix anymore or
|
* we have a less significant route.
|
* we have a less significant route.
|
* we've to insert an intermediate node on the list
|
* we've to insert an intermediate node on the list
|
* this new node will point to the one we need to create
|
* this new node will point to the one we need to create
|
* and the current
|
* and the current
|
*/
|
*/
|
|
|
pn = fn->parent;
|
pn = fn->parent;
|
|
|
/* find 1st bit in difference between the 2 addrs.
|
/* find 1st bit in difference between the 2 addrs.
|
|
|
See comment in addr_diff: bit may be an invalid value,
|
See comment in addr_diff: bit may be an invalid value,
|
but if it is >= plen, the value is ignored in any case.
|
but if it is >= plen, the value is ignored in any case.
|
*/
|
*/
|
|
|
bit = addr_diff(addr, &key->addr, addrlen);
|
bit = addr_diff(addr, &key->addr, addrlen);
|
|
|
/*
|
/*
|
* (intermediate)[in]
|
* (intermediate)[in]
|
* / \
|
* / \
|
* (new leaf node)[ln] (old node)[fn]
|
* (new leaf node)[ln] (old node)[fn]
|
*/
|
*/
|
if (plen > bit) {
|
if (plen > bit) {
|
in = node_alloc();
|
in = node_alloc();
|
ln = node_alloc();
|
ln = node_alloc();
|
|
|
if (in == NULL || ln == NULL) {
|
if (in == NULL || ln == NULL) {
|
if (in)
|
if (in)
|
node_free(in);
|
node_free(in);
|
if (ln)
|
if (ln)
|
node_free(ln);
|
node_free(ln);
|
return NULL;
|
return NULL;
|
}
|
}
|
|
|
/*
|
/*
|
* new intermediate node.
|
* new intermediate node.
|
* RTN_RTINFO will
|
* RTN_RTINFO will
|
* be off since that an address that chooses one of
|
* be off since that an address that chooses one of
|
* the branches would not match less specific routes
|
* the branches would not match less specific routes
|
* in the other branch
|
* in the other branch
|
*/
|
*/
|
|
|
in->fn_bit = bit;
|
in->fn_bit = bit;
|
|
|
in->parent = pn;
|
in->parent = pn;
|
in->leaf = fn->leaf;
|
in->leaf = fn->leaf;
|
atomic_inc(&in->leaf->rt6i_ref);
|
atomic_inc(&in->leaf->rt6i_ref);
|
|
|
in->fn_sernum = sernum;
|
in->fn_sernum = sernum;
|
|
|
/* update parent pointer */
|
/* update parent pointer */
|
if (dir)
|
if (dir)
|
pn->right = in;
|
pn->right = in;
|
else
|
else
|
pn->left = in;
|
pn->left = in;
|
|
|
ln->fn_bit = plen;
|
ln->fn_bit = plen;
|
|
|
ln->parent = in;
|
ln->parent = in;
|
fn->parent = in;
|
fn->parent = in;
|
|
|
ln->fn_sernum = sernum;
|
ln->fn_sernum = sernum;
|
|
|
if (addr_bit_set(addr, bit)) {
|
if (addr_bit_set(addr, bit)) {
|
in->right = ln;
|
in->right = ln;
|
in->left = fn;
|
in->left = fn;
|
} else {
|
} else {
|
in->left = ln;
|
in->left = ln;
|
in->right = fn;
|
in->right = fn;
|
}
|
}
|
} else { /* plen <= bit */
|
} else { /* plen <= bit */
|
|
|
/*
|
/*
|
* (new leaf node)[ln]
|
* (new leaf node)[ln]
|
* / \
|
* / \
|
* (old node)[fn] NULL
|
* (old node)[fn] NULL
|
*/
|
*/
|
|
|
ln = node_alloc();
|
ln = node_alloc();
|
|
|
if (ln == NULL)
|
if (ln == NULL)
|
return NULL;
|
return NULL;
|
|
|
ln->fn_bit = plen;
|
ln->fn_bit = plen;
|
|
|
ln->parent = pn;
|
ln->parent = pn;
|
|
|
ln->fn_sernum = sernum;
|
ln->fn_sernum = sernum;
|
|
|
if (dir)
|
if (dir)
|
pn->right = ln;
|
pn->right = ln;
|
else
|
else
|
pn->left = ln;
|
pn->left = ln;
|
|
|
if (addr_bit_set(&key->addr, plen))
|
if (addr_bit_set(&key->addr, plen))
|
ln->right = fn;
|
ln->right = fn;
|
else
|
else
|
ln->left = fn;
|
ln->left = fn;
|
|
|
fn->parent = ln;
|
fn->parent = ln;
|
}
|
}
|
return ln;
|
return ln;
|
}
|
}
|
|
|
/*
|
/*
|
* Insert routing information in a node.
|
* Insert routing information in a node.
|
*/
|
*/
|
|
|
static int fib6_add_rt2node(struct fib6_node *fn, struct rt6_info *rt,
|
static int fib6_add_rt2node(struct fib6_node *fn, struct rt6_info *rt,
|
struct nlmsghdr *nlh)
|
struct nlmsghdr *nlh)
|
{
|
{
|
struct rt6_info *iter = NULL;
|
struct rt6_info *iter = NULL;
|
struct rt6_info **ins;
|
struct rt6_info **ins;
|
|
|
ins = &fn->leaf;
|
ins = &fn->leaf;
|
|
|
if (fn->fn_flags&RTN_TL_ROOT &&
|
if (fn->fn_flags&RTN_TL_ROOT &&
|
fn->leaf == &ip6_null_entry &&
|
fn->leaf == &ip6_null_entry &&
|
!(rt->rt6i_flags & (RTF_DEFAULT | RTF_ADDRCONF | RTF_ALLONLINK)) ){
|
!(rt->rt6i_flags & (RTF_DEFAULT | RTF_ADDRCONF | RTF_ALLONLINK)) ){
|
/*
|
/*
|
* The top fib of ip6 routing table includes ip6_null_entry.
|
* The top fib of ip6 routing table includes ip6_null_entry.
|
*/
|
*/
|
fn->leaf = rt;
|
fn->leaf = rt;
|
rt->u.next = NULL;
|
rt->u.next = NULL;
|
goto out;
|
goto out;
|
}
|
}
|
|
|
for (iter = fn->leaf; iter; iter=iter->u.next) {
|
for (iter = fn->leaf; iter; iter=iter->u.next) {
|
/*
|
/*
|
* Search for duplicates
|
* Search for duplicates
|
*/
|
*/
|
|
|
if (iter->rt6i_metric == rt->rt6i_metric) {
|
if (iter->rt6i_metric == rt->rt6i_metric) {
|
/*
|
/*
|
* Same priority level
|
* Same priority level
|
*/
|
*/
|
|
|
if ((iter->rt6i_dev == rt->rt6i_dev) &&
|
if ((iter->rt6i_dev == rt->rt6i_dev) &&
|
(iter->rt6i_flowr == rt->rt6i_flowr) &&
|
(iter->rt6i_flowr == rt->rt6i_flowr) &&
|
(ipv6_addr_cmp(&iter->rt6i_gateway,
|
(ipv6_addr_cmp(&iter->rt6i_gateway,
|
&rt->rt6i_gateway) == 0)) {
|
&rt->rt6i_gateway) == 0)) {
|
if (!(iter->rt6i_flags&RTF_EXPIRES))
|
if (!(iter->rt6i_flags&RTF_EXPIRES))
|
return -EEXIST;
|
return -EEXIST;
|
iter->rt6i_expires = rt->rt6i_expires;
|
iter->rt6i_expires = rt->rt6i_expires;
|
if (!(rt->rt6i_flags&RTF_EXPIRES)) {
|
if (!(rt->rt6i_flags&RTF_EXPIRES)) {
|
iter->rt6i_flags &= ~RTF_EXPIRES;
|
iter->rt6i_flags &= ~RTF_EXPIRES;
|
iter->rt6i_expires = 0;
|
iter->rt6i_expires = 0;
|
}
|
}
|
return -EEXIST;
|
return -EEXIST;
|
}
|
}
|
}
|
}
|
|
|
if (iter->rt6i_metric > rt->rt6i_metric)
|
if (iter->rt6i_metric > rt->rt6i_metric)
|
break;
|
break;
|
|
|
ins = &iter->u.next;
|
ins = &iter->u.next;
|
}
|
}
|
|
|
/*
|
/*
|
* insert node
|
* insert node
|
*/
|
*/
|
|
|
out:
|
out:
|
rt->u.next = iter;
|
rt->u.next = iter;
|
*ins = rt;
|
*ins = rt;
|
rt->rt6i_node = fn;
|
rt->rt6i_node = fn;
|
atomic_inc(&rt->rt6i_ref);
|
atomic_inc(&rt->rt6i_ref);
|
inet6_rt_notify(RTM_NEWROUTE, rt, nlh);
|
inet6_rt_notify(RTM_NEWROUTE, rt, nlh);
|
rt6_stats.fib_rt_entries++;
|
rt6_stats.fib_rt_entries++;
|
|
|
if ((fn->fn_flags & RTN_RTINFO) == 0) {
|
if ((fn->fn_flags & RTN_RTINFO) == 0) {
|
rt6_stats.fib_route_nodes++;
|
rt6_stats.fib_route_nodes++;
|
fn->fn_flags |= RTN_RTINFO;
|
fn->fn_flags |= RTN_RTINFO;
|
}
|
}
|
|
|
return 0;
|
return 0;
|
}
|
}
|
|
|
static __inline__ void fib6_start_gc(struct rt6_info *rt)
|
static __inline__ void fib6_start_gc(struct rt6_info *rt)
|
{
|
{
|
if (ip6_fib_timer.expires == 0 &&
|
if (ip6_fib_timer.expires == 0 &&
|
(rt->rt6i_flags & (RTF_EXPIRES|RTF_CACHE)))
|
(rt->rt6i_flags & (RTF_EXPIRES|RTF_CACHE)))
|
mod_timer(&ip6_fib_timer, jiffies + ip6_rt_gc_interval);
|
mod_timer(&ip6_fib_timer, jiffies + ip6_rt_gc_interval);
|
}
|
}
|
|
|
/*
|
/*
|
* Add routing information to the routing tree.
|
* Add routing information to the routing tree.
|
* <destination addr>/<source addr>
|
* <destination addr>/<source addr>
|
* with source addr info in sub-trees
|
* with source addr info in sub-trees
|
*/
|
*/
|
|
|
int fib6_add(struct fib6_node *root, struct rt6_info *rt, struct nlmsghdr *nlh)
|
int fib6_add(struct fib6_node *root, struct rt6_info *rt, struct nlmsghdr *nlh)
|
{
|
{
|
struct fib6_node *fn;
|
struct fib6_node *fn;
|
int err = -ENOMEM;
|
int err = -ENOMEM;
|
|
|
fn = fib6_add_1(root, &rt->rt6i_dst.addr, sizeof(struct in6_addr),
|
fn = fib6_add_1(root, &rt->rt6i_dst.addr, sizeof(struct in6_addr),
|
rt->rt6i_dst.plen, (u8*) &rt->rt6i_dst - (u8*) rt);
|
rt->rt6i_dst.plen, (u8*) &rt->rt6i_dst - (u8*) rt);
|
|
|
if (fn == NULL)
|
if (fn == NULL)
|
goto out;
|
goto out;
|
|
|
#ifdef CONFIG_IPV6_SUBTREES
|
#ifdef CONFIG_IPV6_SUBTREES
|
if (rt->rt6i_src.plen) {
|
if (rt->rt6i_src.plen) {
|
struct fib6_node *sn;
|
struct fib6_node *sn;
|
|
|
if (fn->subtree == NULL) {
|
if (fn->subtree == NULL) {
|
struct fib6_node *sfn;
|
struct fib6_node *sfn;
|
|
|
/*
|
/*
|
* Create subtree.
|
* Create subtree.
|
*
|
*
|
* fn[main tree]
|
* fn[main tree]
|
* |
|
* |
|
* sfn[subtree root]
|
* sfn[subtree root]
|
* \
|
* \
|
* sn[new leaf node]
|
* sn[new leaf node]
|
*/
|
*/
|
|
|
/* Create subtree root node */
|
/* Create subtree root node */
|
sfn = node_alloc();
|
sfn = node_alloc();
|
if (sfn == NULL)
|
if (sfn == NULL)
|
goto st_failure;
|
goto st_failure;
|
|
|
sfn->leaf = &ip6_null_entry;
|
sfn->leaf = &ip6_null_entry;
|
atomic_inc(&ip6_null_entry.rt6i_ref);
|
atomic_inc(&ip6_null_entry.rt6i_ref);
|
sfn->fn_flags = RTN_ROOT;
|
sfn->fn_flags = RTN_ROOT;
|
sfn->fn_sernum = fib6_new_sernum();
|
sfn->fn_sernum = fib6_new_sernum();
|
|
|
/* Now add the first leaf node to new subtree */
|
/* Now add the first leaf node to new subtree */
|
|
|
sn = fib6_add_1(sfn, &rt->rt6i_src.addr,
|
sn = fib6_add_1(sfn, &rt->rt6i_src.addr,
|
sizeof(struct in6_addr), rt->rt6i_src.plen,
|
sizeof(struct in6_addr), rt->rt6i_src.plen,
|
(u8*) &rt->rt6i_src - (u8*) rt);
|
(u8*) &rt->rt6i_src - (u8*) rt);
|
|
|
if (sn == NULL) {
|
if (sn == NULL) {
|
/* If it is failed, discard just allocated
|
/* If it is failed, discard just allocated
|
root, and then (in st_failure) stale node
|
root, and then (in st_failure) stale node
|
in main tree.
|
in main tree.
|
*/
|
*/
|
node_free(sfn);
|
node_free(sfn);
|
goto st_failure;
|
goto st_failure;
|
}
|
}
|
|
|
/* Now link new subtree to main tree */
|
/* Now link new subtree to main tree */
|
sfn->parent = fn;
|
sfn->parent = fn;
|
fn->subtree = sfn;
|
fn->subtree = sfn;
|
if (fn->leaf == NULL) {
|
if (fn->leaf == NULL) {
|
fn->leaf = rt;
|
fn->leaf = rt;
|
atomic_inc(&rt->rt6i_ref);
|
atomic_inc(&rt->rt6i_ref);
|
}
|
}
|
} else {
|
} else {
|
sn = fib6_add_1(fn->subtree, &rt->rt6i_src.addr,
|
sn = fib6_add_1(fn->subtree, &rt->rt6i_src.addr,
|
sizeof(struct in6_addr), rt->rt6i_src.plen,
|
sizeof(struct in6_addr), rt->rt6i_src.plen,
|
(u8*) &rt->rt6i_src - (u8*) rt);
|
(u8*) &rt->rt6i_src - (u8*) rt);
|
|
|
if (sn == NULL)
|
if (sn == NULL)
|
goto st_failure;
|
goto st_failure;
|
}
|
}
|
|
|
fn = sn;
|
fn = sn;
|
}
|
}
|
#endif
|
#endif
|
|
|
err = fib6_add_rt2node(fn, rt, nlh);
|
err = fib6_add_rt2node(fn, rt, nlh);
|
|
|
if (err == 0) {
|
if (err == 0) {
|
fib6_start_gc(rt);
|
fib6_start_gc(rt);
|
if (!(rt->rt6i_flags&RTF_CACHE))
|
if (!(rt->rt6i_flags&RTF_CACHE))
|
fib6_prune_clones(fn, rt);
|
fib6_prune_clones(fn, rt);
|
}
|
}
|
|
|
out:
|
out:
|
if (err)
|
if (err)
|
dst_free(&rt->u.dst);
|
dst_free(&rt->u.dst);
|
return err;
|
return err;
|
|
|
#ifdef CONFIG_IPV6_SUBTREES
|
#ifdef CONFIG_IPV6_SUBTREES
|
/* Subtree creation failed, probably main tree node
|
/* Subtree creation failed, probably main tree node
|
is orphan. If it is, shoot it.
|
is orphan. If it is, shoot it.
|
*/
|
*/
|
st_failure:
|
st_failure:
|
if (fn && !(fn->fn_flags&RTN_RTINFO|RTN_ROOT))
|
if (fn && !(fn->fn_flags&RTN_RTINFO|RTN_ROOT))
|
fib_repair_tree(fn);
|
fib_repair_tree(fn);
|
dst_free(&rt->u.dst);
|
dst_free(&rt->u.dst);
|
return err;
|
return err;
|
#endif
|
#endif
|
}
|
}
|
|
|
/*
|
/*
|
* Routing tree lookup
|
* Routing tree lookup
|
*
|
*
|
*/
|
*/
|
|
|
struct lookup_args {
|
struct lookup_args {
|
int offset; /* key offset on rt6_info */
|
int offset; /* key offset on rt6_info */
|
struct in6_addr *addr; /* search key */
|
struct in6_addr *addr; /* search key */
|
};
|
};
|
|
|
static struct fib6_node * fib6_lookup_1(struct fib6_node *root,
|
static struct fib6_node * fib6_lookup_1(struct fib6_node *root,
|
struct lookup_args *args)
|
struct lookup_args *args)
|
{
|
{
|
struct fib6_node *fn;
|
struct fib6_node *fn;
|
int dir;
|
int dir;
|
|
|
/*
|
/*
|
* Descend on a tree
|
* Descend on a tree
|
*/
|
*/
|
|
|
fn = root;
|
fn = root;
|
|
|
for (;;) {
|
for (;;) {
|
struct fib6_node *next;
|
struct fib6_node *next;
|
|
|
dir = addr_bit_set(args->addr, fn->fn_bit);
|
dir = addr_bit_set(args->addr, fn->fn_bit);
|
|
|
next = dir ? fn->right : fn->left;
|
next = dir ? fn->right : fn->left;
|
|
|
if (next) {
|
if (next) {
|
fn = next;
|
fn = next;
|
continue;
|
continue;
|
}
|
}
|
|
|
break;
|
break;
|
}
|
}
|
|
|
while ((fn->fn_flags & RTN_ROOT) == 0) {
|
while ((fn->fn_flags & RTN_ROOT) == 0) {
|
#ifdef CONFIG_IPV6_SUBTREES
|
#ifdef CONFIG_IPV6_SUBTREES
|
if (fn->subtree) {
|
if (fn->subtree) {
|
struct fib6_node *st;
|
struct fib6_node *st;
|
struct lookup_args *narg;
|
struct lookup_args *narg;
|
|
|
narg = args + 1;
|
narg = args + 1;
|
|
|
if (narg->addr) {
|
if (narg->addr) {
|
st = fib6_lookup_1(fn->subtree, narg);
|
st = fib6_lookup_1(fn->subtree, narg);
|
|
|
if (st && !(st->fn_flags & RTN_ROOT))
|
if (st && !(st->fn_flags & RTN_ROOT))
|
return st;
|
return st;
|
}
|
}
|
}
|
}
|
#endif
|
#endif
|
|
|
if (fn->fn_flags & RTN_RTINFO) {
|
if (fn->fn_flags & RTN_RTINFO) {
|
struct rt6key *key;
|
struct rt6key *key;
|
|
|
key = (struct rt6key *) ((u8 *) fn->leaf +
|
key = (struct rt6key *) ((u8 *) fn->leaf +
|
args->offset);
|
args->offset);
|
|
|
if (addr_match(&key->addr, args->addr, key->plen))
|
if (addr_match(&key->addr, args->addr, key->plen))
|
return fn;
|
return fn;
|
}
|
}
|
|
|
fn = fn->parent;
|
fn = fn->parent;
|
}
|
}
|
|
|
return NULL;
|
return NULL;
|
}
|
}
|
|
|
struct fib6_node * fib6_lookup(struct fib6_node *root, struct in6_addr *daddr,
|
struct fib6_node * fib6_lookup(struct fib6_node *root, struct in6_addr *daddr,
|
struct in6_addr *saddr)
|
struct in6_addr *saddr)
|
{
|
{
|
struct lookup_args args[2];
|
struct lookup_args args[2];
|
struct rt6_info *rt = NULL;
|
struct rt6_info *rt = NULL;
|
struct fib6_node *fn;
|
struct fib6_node *fn;
|
|
|
args[0].offset = (u8*) &rt->rt6i_dst - (u8*) rt;
|
args[0].offset = (u8*) &rt->rt6i_dst - (u8*) rt;
|
args[0].addr = daddr;
|
args[0].addr = daddr;
|
|
|
#ifdef CONFIG_IPV6_SUBTREES
|
#ifdef CONFIG_IPV6_SUBTREES
|
args[1].offset = (u8*) &rt->rt6i_src - (u8*) rt;
|
args[1].offset = (u8*) &rt->rt6i_src - (u8*) rt;
|
args[1].addr = saddr;
|
args[1].addr = saddr;
|
#endif
|
#endif
|
|
|
fn = fib6_lookup_1(root, args);
|
fn = fib6_lookup_1(root, args);
|
|
|
if (fn == NULL || fn->fn_flags & RTN_TL_ROOT)
|
if (fn == NULL || fn->fn_flags & RTN_TL_ROOT)
|
fn = root;
|
fn = root;
|
|
|
return fn;
|
return fn;
|
}
|
}
|
|
|
/*
|
/*
|
* Get node with sepciafied destination prefix (and source prefix,
|
* Get node with sepciafied destination prefix (and source prefix,
|
* if subtrees are used)
|
* if subtrees are used)
|
*/
|
*/
|
|
|
|
|
static struct fib6_node * fib6_locate_1(struct fib6_node *root,
|
static struct fib6_node * fib6_locate_1(struct fib6_node *root,
|
struct in6_addr *addr,
|
struct in6_addr *addr,
|
int plen, int offset)
|
int plen, int offset)
|
{
|
{
|
struct fib6_node *fn;
|
struct fib6_node *fn;
|
|
|
for (fn = root; fn ; ) {
|
for (fn = root; fn ; ) {
|
struct rt6key *key = (struct rt6key *)((u8 *)fn->leaf + offset);
|
struct rt6key *key = (struct rt6key *)((u8 *)fn->leaf + offset);
|
|
|
/*
|
/*
|
* Prefix match
|
* Prefix match
|
*/
|
*/
|
if (plen < fn->fn_bit ||
|
if (plen < fn->fn_bit ||
|
!addr_match(&key->addr, addr, fn->fn_bit))
|
!addr_match(&key->addr, addr, fn->fn_bit))
|
return NULL;
|
return NULL;
|
|
|
if (plen == fn->fn_bit)
|
if (plen == fn->fn_bit)
|
return fn;
|
return fn;
|
|
|
/*
|
/*
|
* We have more bits to go
|
* We have more bits to go
|
*/
|
*/
|
if (addr_bit_set(addr, fn->fn_bit))
|
if (addr_bit_set(addr, fn->fn_bit))
|
fn = fn->right;
|
fn = fn->right;
|
else
|
else
|
fn = fn->left;
|
fn = fn->left;
|
}
|
}
|
return NULL;
|
return NULL;
|
}
|
}
|
|
|
struct fib6_node * fib6_locate(struct fib6_node *root,
|
struct fib6_node * fib6_locate(struct fib6_node *root,
|
struct in6_addr *daddr, int dst_len,
|
struct in6_addr *daddr, int dst_len,
|
struct in6_addr *saddr, int src_len)
|
struct in6_addr *saddr, int src_len)
|
{
|
{
|
struct rt6_info *rt = NULL;
|
struct rt6_info *rt = NULL;
|
struct fib6_node *fn;
|
struct fib6_node *fn;
|
|
|
fn = fib6_locate_1(root, daddr, dst_len,
|
fn = fib6_locate_1(root, daddr, dst_len,
|
(u8*) &rt->rt6i_dst - (u8*) rt);
|
(u8*) &rt->rt6i_dst - (u8*) rt);
|
|
|
#ifdef CONFIG_IPV6_SUBTREES
|
#ifdef CONFIG_IPV6_SUBTREES
|
if (src_len) {
|
if (src_len) {
|
BUG_TRAP(saddr!=NULL);
|
BUG_TRAP(saddr!=NULL);
|
if (fn == NULL)
|
if (fn == NULL)
|
fn = fn->subtree;
|
fn = fn->subtree;
|
if (fn)
|
if (fn)
|
fn = fib6_locate_1(fn, saddr, src_len,
|
fn = fib6_locate_1(fn, saddr, src_len,
|
(u8*) &rt->rt6i_src - (u8*) rt);
|
(u8*) &rt->rt6i_src - (u8*) rt);
|
}
|
}
|
#endif
|
#endif
|
|
|
if (fn && fn->fn_flags&RTN_RTINFO)
|
if (fn && fn->fn_flags&RTN_RTINFO)
|
return fn;
|
return fn;
|
|
|
return NULL;
|
return NULL;
|
}
|
}
|
|
|
|
|
/*
|
/*
|
* Deletion
|
* Deletion
|
*
|
*
|
*/
|
*/
|
|
|
static struct rt6_info * fib6_find_prefix(struct fib6_node *fn)
|
static struct rt6_info * fib6_find_prefix(struct fib6_node *fn)
|
{
|
{
|
if (fn->fn_flags&RTN_ROOT)
|
if (fn->fn_flags&RTN_ROOT)
|
return &ip6_null_entry;
|
return &ip6_null_entry;
|
|
|
while(fn) {
|
while(fn) {
|
if(fn->left)
|
if(fn->left)
|
return fn->left->leaf;
|
return fn->left->leaf;
|
|
|
if(fn->right)
|
if(fn->right)
|
return fn->right->leaf;
|
return fn->right->leaf;
|
|
|
fn = SUBTREE(fn);
|
fn = SUBTREE(fn);
|
}
|
}
|
return NULL;
|
return NULL;
|
}
|
}
|
|
|
/*
|
/*
|
* Called to trim the tree of intermediate nodes when possible. "fn"
|
* Called to trim the tree of intermediate nodes when possible. "fn"
|
* is the node we want to try and remove.
|
* is the node we want to try and remove.
|
*/
|
*/
|
|
|
static struct fib6_node * fib6_repair_tree(struct fib6_node *fn)
|
static struct fib6_node * fib6_repair_tree(struct fib6_node *fn)
|
{
|
{
|
int children;
|
int children;
|
int nstate;
|
int nstate;
|
struct fib6_node *child, *pn;
|
struct fib6_node *child, *pn;
|
struct fib6_walker_t *w;
|
struct fib6_walker_t *w;
|
int iter = 0;
|
int iter = 0;
|
|
|
for (;;) {
|
for (;;) {
|
RT6_TRACE("fixing tree: plen=%d iter=%d\n", fn->fn_bit, iter);
|
RT6_TRACE("fixing tree: plen=%d iter=%d\n", fn->fn_bit, iter);
|
iter++;
|
iter++;
|
|
|
BUG_TRAP(!(fn->fn_flags&RTN_RTINFO));
|
BUG_TRAP(!(fn->fn_flags&RTN_RTINFO));
|
BUG_TRAP(!(fn->fn_flags&RTN_TL_ROOT));
|
BUG_TRAP(!(fn->fn_flags&RTN_TL_ROOT));
|
BUG_TRAP(fn->leaf==NULL);
|
BUG_TRAP(fn->leaf==NULL);
|
|
|
children = 0;
|
children = 0;
|
child = NULL;
|
child = NULL;
|
if (fn->right) child = fn->right, children |= 1;
|
if (fn->right) child = fn->right, children |= 1;
|
if (fn->left) child = fn->left, children |= 2;
|
if (fn->left) child = fn->left, children |= 2;
|
|
|
if (children == 3 || SUBTREE(fn)
|
if (children == 3 || SUBTREE(fn)
|
#ifdef CONFIG_IPV6_SUBTREES
|
#ifdef CONFIG_IPV6_SUBTREES
|
/* Subtree root (i.e. fn) may have one child */
|
/* Subtree root (i.e. fn) may have one child */
|
|| (children && fn->fn_flags&RTN_ROOT)
|
|| (children && fn->fn_flags&RTN_ROOT)
|
#endif
|
#endif
|
) {
|
) {
|
fn->leaf = fib6_find_prefix(fn);
|
fn->leaf = fib6_find_prefix(fn);
|
#if RT6_DEBUG >= 2
|
#if RT6_DEBUG >= 2
|
if (fn->leaf==NULL) {
|
if (fn->leaf==NULL) {
|
BUG_TRAP(fn->leaf);
|
BUG_TRAP(fn->leaf);
|
fn->leaf = &ip6_null_entry;
|
fn->leaf = &ip6_null_entry;
|
}
|
}
|
#endif
|
#endif
|
atomic_inc(&fn->leaf->rt6i_ref);
|
atomic_inc(&fn->leaf->rt6i_ref);
|
return fn->parent;
|
return fn->parent;
|
}
|
}
|
|
|
pn = fn->parent;
|
pn = fn->parent;
|
#ifdef CONFIG_IPV6_SUBTREES
|
#ifdef CONFIG_IPV6_SUBTREES
|
if (SUBTREE(pn) == fn) {
|
if (SUBTREE(pn) == fn) {
|
BUG_TRAP(fn->fn_flags&RTN_ROOT);
|
BUG_TRAP(fn->fn_flags&RTN_ROOT);
|
SUBTREE(pn) = NULL;
|
SUBTREE(pn) = NULL;
|
nstate = FWS_L;
|
nstate = FWS_L;
|
} else {
|
} else {
|
BUG_TRAP(!(fn->fn_flags&RTN_ROOT));
|
BUG_TRAP(!(fn->fn_flags&RTN_ROOT));
|
#endif
|
#endif
|
if (pn->right == fn) pn->right = child;
|
if (pn->right == fn) pn->right = child;
|
else if (pn->left == fn) pn->left = child;
|
else if (pn->left == fn) pn->left = child;
|
#if RT6_DEBUG >= 2
|
#if RT6_DEBUG >= 2
|
else BUG_TRAP(0);
|
else BUG_TRAP(0);
|
#endif
|
#endif
|
if (child)
|
if (child)
|
child->parent = pn;
|
child->parent = pn;
|
nstate = FWS_R;
|
nstate = FWS_R;
|
#ifdef CONFIG_IPV6_SUBTREES
|
#ifdef CONFIG_IPV6_SUBTREES
|
}
|
}
|
#endif
|
#endif
|
|
|
read_lock(&fib6_walker_lock);
|
read_lock(&fib6_walker_lock);
|
FOR_WALKERS(w) {
|
FOR_WALKERS(w) {
|
if (child == NULL) {
|
if (child == NULL) {
|
if (w->root == fn) {
|
if (w->root == fn) {
|
w->root = w->node = NULL;
|
w->root = w->node = NULL;
|
RT6_TRACE("W %p adjusted by delroot 1\n", w);
|
RT6_TRACE("W %p adjusted by delroot 1\n", w);
|
} else if (w->node == fn) {
|
} else if (w->node == fn) {
|
RT6_TRACE("W %p adjusted by delnode 1, s=%d/%d\n", w, w->state, nstate);
|
RT6_TRACE("W %p adjusted by delnode 1, s=%d/%d\n", w, w->state, nstate);
|
w->node = pn;
|
w->node = pn;
|
w->state = nstate;
|
w->state = nstate;
|
}
|
}
|
} else {
|
} else {
|
if (w->root == fn) {
|
if (w->root == fn) {
|
w->root = child;
|
w->root = child;
|
RT6_TRACE("W %p adjusted by delroot 2\n", w);
|
RT6_TRACE("W %p adjusted by delroot 2\n", w);
|
}
|
}
|
if (w->node == fn) {
|
if (w->node == fn) {
|
w->node = child;
|
w->node = child;
|
if (children&2) {
|
if (children&2) {
|
RT6_TRACE("W %p adjusted by delnode 2, s=%d\n", w, w->state);
|
RT6_TRACE("W %p adjusted by delnode 2, s=%d\n", w, w->state);
|
w->state = w->state>=FWS_R ? FWS_U : FWS_INIT;
|
w->state = w->state>=FWS_R ? FWS_U : FWS_INIT;
|
} else {
|
} else {
|
RT6_TRACE("W %p adjusted by delnode 2, s=%d\n", w, w->state);
|
RT6_TRACE("W %p adjusted by delnode 2, s=%d\n", w, w->state);
|
w->state = w->state>=FWS_C ? FWS_U : FWS_INIT;
|
w->state = w->state>=FWS_C ? FWS_U : FWS_INIT;
|
}
|
}
|
}
|
}
|
}
|
}
|
}
|
}
|
read_unlock(&fib6_walker_lock);
|
read_unlock(&fib6_walker_lock);
|
|
|
node_free(fn);
|
node_free(fn);
|
if (pn->fn_flags&RTN_RTINFO || SUBTREE(pn))
|
if (pn->fn_flags&RTN_RTINFO || SUBTREE(pn))
|
return pn;
|
return pn;
|
|
|
rt6_release(pn->leaf);
|
rt6_release(pn->leaf);
|
pn->leaf = NULL;
|
pn->leaf = NULL;
|
fn = pn;
|
fn = pn;
|
}
|
}
|
}
|
}
|
|
|
static void fib6_del_route(struct fib6_node *fn, struct rt6_info **rtp,
|
static void fib6_del_route(struct fib6_node *fn, struct rt6_info **rtp,
|
struct nlmsghdr *nlh)
|
struct nlmsghdr *nlh)
|
{
|
{
|
struct fib6_walker_t *w;
|
struct fib6_walker_t *w;
|
struct rt6_info *rt = *rtp;
|
struct rt6_info *rt = *rtp;
|
|
|
RT6_TRACE("fib6_del_route\n");
|
RT6_TRACE("fib6_del_route\n");
|
|
|
/* Unlink it */
|
/* Unlink it */
|
*rtp = rt->u.next;
|
*rtp = rt->u.next;
|
rt->rt6i_node = NULL;
|
rt->rt6i_node = NULL;
|
rt6_stats.fib_rt_entries--;
|
rt6_stats.fib_rt_entries--;
|
|
|
/* Adjust walkers */
|
/* Adjust walkers */
|
read_lock(&fib6_walker_lock);
|
read_lock(&fib6_walker_lock);
|
FOR_WALKERS(w) {
|
FOR_WALKERS(w) {
|
if (w->state == FWS_C && w->leaf == rt) {
|
if (w->state == FWS_C && w->leaf == rt) {
|
RT6_TRACE("walker %p adjusted by delroute\n", w);
|
RT6_TRACE("walker %p adjusted by delroute\n", w);
|
w->leaf = rt->u.next;
|
w->leaf = rt->u.next;
|
if (w->leaf == NULL)
|
if (w->leaf == NULL)
|
w->state = FWS_U;
|
w->state = FWS_U;
|
}
|
}
|
}
|
}
|
read_unlock(&fib6_walker_lock);
|
read_unlock(&fib6_walker_lock);
|
|
|
rt->u.next = NULL;
|
rt->u.next = NULL;
|
|
|
if (fn->leaf == NULL && fn->fn_flags&RTN_TL_ROOT)
|
if (fn->leaf == NULL && fn->fn_flags&RTN_TL_ROOT)
|
fn->leaf = &ip6_null_entry;
|
fn->leaf = &ip6_null_entry;
|
|
|
/* If it was last route, expunge its radix tree node */
|
/* If it was last route, expunge its radix tree node */
|
if (fn->leaf == NULL) {
|
if (fn->leaf == NULL) {
|
fn->fn_flags &= ~RTN_RTINFO;
|
fn->fn_flags &= ~RTN_RTINFO;
|
rt6_stats.fib_route_nodes--;
|
rt6_stats.fib_route_nodes--;
|
fn = fib6_repair_tree(fn);
|
fn = fib6_repair_tree(fn);
|
}
|
}
|
|
|
if (atomic_read(&rt->rt6i_ref) != 1) {
|
if (atomic_read(&rt->rt6i_ref) != 1) {
|
/* This route is used as dummy address holder in some split
|
/* This route is used as dummy address holder in some split
|
* nodes. It is not leaked, but it still holds other resources,
|
* nodes. It is not leaked, but it still holds other resources,
|
* which must be released in time. So, scan ascendant nodes
|
* which must be released in time. So, scan ascendant nodes
|
* and replace dummy references to this route with references
|
* and replace dummy references to this route with references
|
* to still alive ones.
|
* to still alive ones.
|
*/
|
*/
|
while (fn) {
|
while (fn) {
|
if (!(fn->fn_flags&RTN_RTINFO) && fn->leaf == rt) {
|
if (!(fn->fn_flags&RTN_RTINFO) && fn->leaf == rt) {
|
fn->leaf = fib6_find_prefix(fn);
|
fn->leaf = fib6_find_prefix(fn);
|
atomic_inc(&fn->leaf->rt6i_ref);
|
atomic_inc(&fn->leaf->rt6i_ref);
|
rt6_release(rt);
|
rt6_release(rt);
|
}
|
}
|
fn = fn->parent;
|
fn = fn->parent;
|
}
|
}
|
/* No more references are possiible at this point. */
|
/* No more references are possiible at this point. */
|
if (atomic_read(&rt->rt6i_ref) != 1) BUG();
|
if (atomic_read(&rt->rt6i_ref) != 1) BUG();
|
}
|
}
|
|
|
inet6_rt_notify(RTM_DELROUTE, rt, nlh);
|
inet6_rt_notify(RTM_DELROUTE, rt, nlh);
|
rt6_release(rt);
|
rt6_release(rt);
|
}
|
}
|
|
|
int fib6_del(struct rt6_info *rt, struct nlmsghdr *nlh)
|
int fib6_del(struct rt6_info *rt, struct nlmsghdr *nlh)
|
{
|
{
|
struct fib6_node *fn = rt->rt6i_node;
|
struct fib6_node *fn = rt->rt6i_node;
|
struct rt6_info **rtp;
|
struct rt6_info **rtp;
|
|
|
#if RT6_DEBUG >= 2
|
#if RT6_DEBUG >= 2
|
if (rt->u.dst.obsolete>0) {
|
if (rt->u.dst.obsolete>0) {
|
BUG_TRAP(fn==NULL);
|
BUG_TRAP(fn==NULL);
|
return -ENOENT;
|
return -ENOENT;
|
}
|
}
|
#endif
|
#endif
|
if (fn == NULL || rt == &ip6_null_entry)
|
if (fn == NULL || rt == &ip6_null_entry)
|
return -ENOENT;
|
return -ENOENT;
|
|
|
BUG_TRAP(fn->fn_flags&RTN_RTINFO);
|
BUG_TRAP(fn->fn_flags&RTN_RTINFO);
|
|
|
if (!(rt->rt6i_flags&RTF_CACHE))
|
if (!(rt->rt6i_flags&RTF_CACHE))
|
fib6_prune_clones(fn, rt);
|
fib6_prune_clones(fn, rt);
|
|
|
/*
|
/*
|
* Walk the leaf entries looking for ourself
|
* Walk the leaf entries looking for ourself
|
*/
|
*/
|
|
|
for (rtp = &fn->leaf; *rtp; rtp = &(*rtp)->u.next) {
|
for (rtp = &fn->leaf; *rtp; rtp = &(*rtp)->u.next) {
|
if (*rtp == rt) {
|
if (*rtp == rt) {
|
fib6_del_route(fn, rtp, nlh);
|
fib6_del_route(fn, rtp, nlh);
|
return 0;
|
return 0;
|
}
|
}
|
}
|
}
|
return -ENOENT;
|
return -ENOENT;
|
}
|
}
|
|
|
/*
|
/*
|
* Tree traversal function.
|
* Tree traversal function.
|
*
|
*
|
* Certainly, it is not interrupt safe.
|
* Certainly, it is not interrupt safe.
|
* However, it is internally reenterable wrt itself and fib6_add/fib6_del.
|
* However, it is internally reenterable wrt itself and fib6_add/fib6_del.
|
* It means, that we can modify tree during walking
|
* It means, that we can modify tree during walking
|
* and use this function for garbage collection, clone pruning,
|
* and use this function for garbage collection, clone pruning,
|
* cleaning tree when a device goes down etc. etc.
|
* cleaning tree when a device goes down etc. etc.
|
*
|
*
|
* It guarantees that every node will be traversed,
|
* It guarantees that every node will be traversed,
|
* and that it will be traversed only once.
|
* and that it will be traversed only once.
|
*
|
*
|
* Callback function w->func may return:
|
* Callback function w->func may return:
|
* 0 -> continue walking.
|
* 0 -> continue walking.
|
* positive value -> walking is suspended (used by tree dumps,
|
* positive value -> walking is suspended (used by tree dumps,
|
* and probably by gc, if it will be split to several slices)
|
* and probably by gc, if it will be split to several slices)
|
* negative value -> terminate walking.
|
* negative value -> terminate walking.
|
*
|
*
|
* The function itself returns:
|
* The function itself returns:
|
* 0 -> walk is complete.
|
* 0 -> walk is complete.
|
* >0 -> walk is incomplete (i.e. suspended)
|
* >0 -> walk is incomplete (i.e. suspended)
|
* <0 -> walk is terminated by an error.
|
* <0 -> walk is terminated by an error.
|
*/
|
*/
|
|
|
int fib6_walk_continue(struct fib6_walker_t *w)
|
int fib6_walk_continue(struct fib6_walker_t *w)
|
{
|
{
|
struct fib6_node *fn, *pn;
|
struct fib6_node *fn, *pn;
|
|
|
for (;;) {
|
for (;;) {
|
fn = w->node;
|
fn = w->node;
|
if (fn == NULL)
|
if (fn == NULL)
|
return 0;
|
return 0;
|
|
|
if (w->prune && fn != w->root &&
|
if (w->prune && fn != w->root &&
|
fn->fn_flags&RTN_RTINFO && w->state < FWS_C) {
|
fn->fn_flags&RTN_RTINFO && w->state < FWS_C) {
|
w->state = FWS_C;
|
w->state = FWS_C;
|
w->leaf = fn->leaf;
|
w->leaf = fn->leaf;
|
}
|
}
|
switch (w->state) {
|
switch (w->state) {
|
#ifdef CONFIG_IPV6_SUBTREES
|
#ifdef CONFIG_IPV6_SUBTREES
|
case FWS_S:
|
case FWS_S:
|
if (SUBTREE(fn)) {
|
if (SUBTREE(fn)) {
|
w->node = SUBTREE(fn);
|
w->node = SUBTREE(fn);
|
continue;
|
continue;
|
}
|
}
|
w->state = FWS_L;
|
w->state = FWS_L;
|
#endif
|
#endif
|
case FWS_L:
|
case FWS_L:
|
if (fn->left) {
|
if (fn->left) {
|
w->node = fn->left;
|
w->node = fn->left;
|
w->state = FWS_INIT;
|
w->state = FWS_INIT;
|
continue;
|
continue;
|
}
|
}
|
w->state = FWS_R;
|
w->state = FWS_R;
|
case FWS_R:
|
case FWS_R:
|
if (fn->right) {
|
if (fn->right) {
|
w->node = fn->right;
|
w->node = fn->right;
|
w->state = FWS_INIT;
|
w->state = FWS_INIT;
|
continue;
|
continue;
|
}
|
}
|
w->state = FWS_C;
|
w->state = FWS_C;
|
w->leaf = fn->leaf;
|
w->leaf = fn->leaf;
|
case FWS_C:
|
case FWS_C:
|
if (w->leaf && fn->fn_flags&RTN_RTINFO) {
|
if (w->leaf && fn->fn_flags&RTN_RTINFO) {
|
int err = w->func(w);
|
int err = w->func(w);
|
if (err)
|
if (err)
|
return err;
|
return err;
|
continue;
|
continue;
|
}
|
}
|
w->state = FWS_U;
|
w->state = FWS_U;
|
case FWS_U:
|
case FWS_U:
|
if (fn == w->root)
|
if (fn == w->root)
|
return 0;
|
return 0;
|
pn = fn->parent;
|
pn = fn->parent;
|
w->node = pn;
|
w->node = pn;
|
#ifdef CONFIG_IPV6_SUBTREES
|
#ifdef CONFIG_IPV6_SUBTREES
|
if (SUBTREE(pn) == fn) {
|
if (SUBTREE(pn) == fn) {
|
BUG_TRAP(fn->fn_flags&RTN_ROOT);
|
BUG_TRAP(fn->fn_flags&RTN_ROOT);
|
w->state = FWS_L;
|
w->state = FWS_L;
|
continue;
|
continue;
|
}
|
}
|
#endif
|
#endif
|
if (pn->left == fn) {
|
if (pn->left == fn) {
|
w->state = FWS_R;
|
w->state = FWS_R;
|
continue;
|
continue;
|
}
|
}
|
if (pn->right == fn) {
|
if (pn->right == fn) {
|
w->state = FWS_C;
|
w->state = FWS_C;
|
w->leaf = w->node->leaf;
|
w->leaf = w->node->leaf;
|
continue;
|
continue;
|
}
|
}
|
#if RT6_DEBUG >= 2
|
#if RT6_DEBUG >= 2
|
BUG_TRAP(0);
|
BUG_TRAP(0);
|
#endif
|
#endif
|
}
|
}
|
}
|
}
|
}
|
}
|
|
|
int fib6_walk(struct fib6_walker_t *w)
|
int fib6_walk(struct fib6_walker_t *w)
|
{
|
{
|
int res;
|
int res;
|
|
|
w->state = FWS_INIT;
|
w->state = FWS_INIT;
|
w->node = w->root;
|
w->node = w->root;
|
|
|
fib6_walker_link(w);
|
fib6_walker_link(w);
|
res = fib6_walk_continue(w);
|
res = fib6_walk_continue(w);
|
if (res <= 0)
|
if (res <= 0)
|
fib6_walker_unlink(w);
|
fib6_walker_unlink(w);
|
return res;
|
return res;
|
}
|
}
|
|
|
static int fib6_clean_node(struct fib6_walker_t *w)
|
static int fib6_clean_node(struct fib6_walker_t *w)
|
{
|
{
|
int res;
|
int res;
|
struct rt6_info *rt;
|
struct rt6_info *rt;
|
struct fib6_cleaner_t *c = (struct fib6_cleaner_t*)w;
|
struct fib6_cleaner_t *c = (struct fib6_cleaner_t*)w;
|
|
|
for (rt = w->leaf; rt; rt = rt->u.next) {
|
for (rt = w->leaf; rt; rt = rt->u.next) {
|
res = c->func(rt, c->arg);
|
res = c->func(rt, c->arg);
|
if (res < 0) {
|
if (res < 0) {
|
w->leaf = rt;
|
w->leaf = rt;
|
res = fib6_del(rt, NULL);
|
res = fib6_del(rt, NULL);
|
if (res) {
|
if (res) {
|
#if RT6_DEBUG >= 2
|
#if RT6_DEBUG >= 2
|
printk(KERN_DEBUG "fib6_clean_node: del failed: rt=%p@%p err=%d\n", rt, rt->rt6i_node, res);
|
printk(KERN_DEBUG "fib6_clean_node: del failed: rt=%p@%p err=%d\n", rt, rt->rt6i_node, res);
|
#endif
|
#endif
|
continue;
|
continue;
|
}
|
}
|
return 0;
|
return 0;
|
}
|
}
|
BUG_TRAP(res==0);
|
BUG_TRAP(res==0);
|
}
|
}
|
w->leaf = rt;
|
w->leaf = rt;
|
return 0;
|
return 0;
|
}
|
}
|
|
|
/*
|
/*
|
* Convenient frontend to tree walker.
|
* Convenient frontend to tree walker.
|
*
|
*
|
* func is called on each route.
|
* func is called on each route.
|
* It may return -1 -> delete this route.
|
* It may return -1 -> delete this route.
|
* 0 -> continue walking
|
* 0 -> continue walking
|
*
|
*
|
* prune==1 -> only immediate children of node (certainly,
|
* prune==1 -> only immediate children of node (certainly,
|
* ignoring pure split nodes) will be scanned.
|
* ignoring pure split nodes) will be scanned.
|
*/
|
*/
|
|
|
void fib6_clean_tree(struct fib6_node *root,
|
void fib6_clean_tree(struct fib6_node *root,
|
int (*func)(struct rt6_info *, void *arg),
|
int (*func)(struct rt6_info *, void *arg),
|
int prune, void *arg)
|
int prune, void *arg)
|
{
|
{
|
struct fib6_cleaner_t c;
|
struct fib6_cleaner_t c;
|
|
|
c.w.root = root;
|
c.w.root = root;
|
c.w.func = fib6_clean_node;
|
c.w.func = fib6_clean_node;
|
c.w.prune = prune;
|
c.w.prune = prune;
|
c.func = func;
|
c.func = func;
|
c.arg = arg;
|
c.arg = arg;
|
|
|
fib6_walk(&c.w);
|
fib6_walk(&c.w);
|
}
|
}
|
|
|
static int fib6_prune_clone(struct rt6_info *rt, void *arg)
|
static int fib6_prune_clone(struct rt6_info *rt, void *arg)
|
{
|
{
|
if (rt->rt6i_flags & RTF_CACHE) {
|
if (rt->rt6i_flags & RTF_CACHE) {
|
RT6_TRACE("pruning clone %p\n", rt);
|
RT6_TRACE("pruning clone %p\n", rt);
|
return -1;
|
return -1;
|
}
|
}
|
|
|
return 0;
|
return 0;
|
}
|
}
|
|
|
static void fib6_prune_clones(struct fib6_node *fn, struct rt6_info *rt)
|
static void fib6_prune_clones(struct fib6_node *fn, struct rt6_info *rt)
|
{
|
{
|
fib6_clean_tree(fn, fib6_prune_clone, 1, rt);
|
fib6_clean_tree(fn, fib6_prune_clone, 1, rt);
|
}
|
}
|
|
|
/*
|
/*
|
* Garbage collection
|
* Garbage collection
|
*/
|
*/
|
|
|
static struct fib6_gc_args
|
static struct fib6_gc_args
|
{
|
{
|
int timeout;
|
int timeout;
|
int more;
|
int more;
|
} gc_args;
|
} gc_args;
|
|
|
static int fib6_age(struct rt6_info *rt, void *arg)
|
static int fib6_age(struct rt6_info *rt, void *arg)
|
{
|
{
|
unsigned long now = jiffies;
|
unsigned long now = jiffies;
|
|
|
/* Age clones. Note, that clones are aged out
|
/* Age clones. Note, that clones are aged out
|
only if they are not in use now.
|
only if they are not in use now.
|
*/
|
*/
|
|
|
/*
|
/*
|
* check addrconf expiration here.
|
* check addrconf expiration here.
|
* They are expired even if they are in use.
|
* They are expired even if they are in use.
|
*/
|
*/
|
|
|
if (rt->rt6i_flags&RTF_EXPIRES && rt->rt6i_expires) {
|
if (rt->rt6i_flags&RTF_EXPIRES && rt->rt6i_expires) {
|
if (time_after(now, rt->rt6i_expires)) {
|
if (time_after(now, rt->rt6i_expires)) {
|
RT6_TRACE("expiring %p\n", rt);
|
RT6_TRACE("expiring %p\n", rt);
|
return -1;
|
return -1;
|
}
|
}
|
gc_args.more++;
|
gc_args.more++;
|
} else if (rt->rt6i_flags & RTF_CACHE) {
|
} else if (rt->rt6i_flags & RTF_CACHE) {
|
if (atomic_read(&rt->u.dst.__refcnt) == 0 &&
|
if (atomic_read(&rt->u.dst.__refcnt) == 0 &&
|
time_after_eq(now, rt->u.dst.lastuse + gc_args.timeout)) {
|
time_after_eq(now, rt->u.dst.lastuse + gc_args.timeout)) {
|
RT6_TRACE("aging clone %p\n", rt);
|
RT6_TRACE("aging clone %p\n", rt);
|
return -1;
|
return -1;
|
}
|
}
|
gc_args.more++;
|
gc_args.more++;
|
}
|
}
|
|
|
return 0;
|
return 0;
|
}
|
}
|
|
|
static spinlock_t fib6_gc_lock = SPIN_LOCK_UNLOCKED;
|
static spinlock_t fib6_gc_lock = SPIN_LOCK_UNLOCKED;
|
|
|
void fib6_run_gc(unsigned long dummy)
|
void fib6_run_gc(unsigned long dummy)
|
{
|
{
|
if (dummy != ~0UL) {
|
if (dummy != ~0UL) {
|
spin_lock_bh(&fib6_gc_lock);
|
spin_lock_bh(&fib6_gc_lock);
|
gc_args.timeout = (int)dummy;
|
gc_args.timeout = (int)dummy;
|
} else {
|
} else {
|
local_bh_disable();
|
local_bh_disable();
|
if (!spin_trylock(&fib6_gc_lock)) {
|
if (!spin_trylock(&fib6_gc_lock)) {
|
mod_timer(&ip6_fib_timer, jiffies + HZ);
|
mod_timer(&ip6_fib_timer, jiffies + HZ);
|
local_bh_enable();
|
local_bh_enable();
|
return;
|
return;
|
}
|
}
|
gc_args.timeout = ip6_rt_gc_interval;
|
gc_args.timeout = ip6_rt_gc_interval;
|
}
|
}
|
gc_args.more = 0;
|
gc_args.more = 0;
|
|
|
|
|
write_lock_bh(&rt6_lock);
|
write_lock_bh(&rt6_lock);
|
fib6_clean_tree(&ip6_routing_table, fib6_age, 0, NULL);
|
fib6_clean_tree(&ip6_routing_table, fib6_age, 0, NULL);
|
write_unlock_bh(&rt6_lock);
|
write_unlock_bh(&rt6_lock);
|
|
|
if (gc_args.more)
|
if (gc_args.more)
|
mod_timer(&ip6_fib_timer, jiffies + ip6_rt_gc_interval);
|
mod_timer(&ip6_fib_timer, jiffies + ip6_rt_gc_interval);
|
else {
|
else {
|
del_timer(&ip6_fib_timer);
|
del_timer(&ip6_fib_timer);
|
ip6_fib_timer.expires = 0;
|
ip6_fib_timer.expires = 0;
|
}
|
}
|
spin_unlock_bh(&fib6_gc_lock);
|
spin_unlock_bh(&fib6_gc_lock);
|
}
|
}
|
|
|
void __init fib6_init(void)
|
void __init fib6_init(void)
|
{
|
{
|
if (!fib6_node_kmem)
|
if (!fib6_node_kmem)
|
fib6_node_kmem = kmem_cache_create("fib6_nodes",
|
fib6_node_kmem = kmem_cache_create("fib6_nodes",
|
sizeof(struct fib6_node),
|
sizeof(struct fib6_node),
|
0, SLAB_HWCACHE_ALIGN,
|
0, SLAB_HWCACHE_ALIGN,
|
NULL, NULL);
|
NULL, NULL);
|
}
|
}
|
|
|
#ifdef MODULE
|
#ifdef MODULE
|
void fib6_gc_cleanup(void)
|
void fib6_gc_cleanup(void)
|
{
|
{
|
del_timer(&ip6_fib_timer);
|
del_timer(&ip6_fib_timer);
|
}
|
}
|
#endif
|
#endif
|
|
|
|
|
|
|
